<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Atif Hussain on Medium]]></title>
        <description><![CDATA[Stories by Atif Hussain on Medium]]></description>
        <link>https://medium.com/@atifhussain_54530?source=rss-8c9f7d2b55fc------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*WrOAFuchh8jstkY9.</url>
            <title>Stories by Atif Hussain on Medium</title>
            <link>https://medium.com/@atifhussain_54530?source=rss-8c9f7d2b55fc------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 15:48:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@atifhussain_54530/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Why Rich Drive?]]></title>
            <link>https://medium.com/@atifhussain_54530/why-rich-drive-a971f1fbf9fb?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/a971f1fbf9fb</guid>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[mobility]]></category>
            <category><![CDATA[commute]]></category>
            <category><![CDATA[whatscoming]]></category>
            <category><![CDATA[public-transportation]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Mon, 05 Oct 2026 05:57:08 GMT</pubDate>
            <atom:updated>2026-10-05T06:00:17.293Z</atom:updated>
            <content:encoded><![CDATA[<p>Why do the rich, who can afford most comforts, still have to drive in congested city traffic? we show 3 gaps, Public Transport can, but is not fulfilling. hint: how food delivery succeeds</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*TQz5NMlAMIgXXEPy.png" /></figure><h3>It isn’t the fare. It’s three gaps that money alone can’t close.</h3><p>Development is usually measured one way: the day a poor family can finally afford a car. There is a better test. The better test is the day someone who can already afford a car, or a driver to drive it, chooses the bus instead.</p><p>That day is not arriving in most Indian metros. This is about why, and what actually closes the gap. Not more rail. Not wider roads. Three specific fixes, aimed at three specific gaps.</p><h3>Bengaluru’s fleet, and who’s really driving</h3><p>Bengaluru’s registered vehicle fleet stood at 1.28 crore as of 30 June 2026. That is 87 lakh two-wheelers and 25.2 lakh private cars, the largest car fleet of any Indian city. Add the two together and they come to nearly 9 in 10 of every vehicle on the city’s rolls.</p><p>This is not a claim about who is voting against the metro, or about distrust. It is simply the scale of who is driving themselves, or being driven, instead of taking a shared seat. Many of these owners are not poor. Affording the fare has never been their problem.</p><h3>Ease, not fare</h3><p>In a metro, once someone can afford to live in the city at all, a bus ticket is rarely the obstacle. Ease is. A commute that costs more time, more waiting, or more uncertainty than driving loses, even when it costs less money.</p><p>Three specific gaps keep ease from winning.</p><h3>Gap one: capacity</h3><p>At the hours people actually travel, there simply aren’t enough buses or shared vehicles on the road. A service that runs fine at noon and fails at 9 AM has, in practice, failed at 9 AM. That is the only time that matters to a commuter.</p><h3>Gap two: the last mile</h3><p>A metro corridor is excellent at what it does. It moves a lot of people fast, in a straight line, between fixed stations. That is real capacity, and nothing here argues against building more of it.</p><p>But a corridor is supply measured in kilometres. A commute is demand measured in doors: your door to their door, at the hour you actually travel. The corridor can be excellent and the trip can still fail, on the stretch between the station and an actual office gate, or a narrow residential lane no large vehicle can enter. That stretch is not a smaller version of the corridor’s job. It is a different job, and it needs a different, smaller vehicle.</p><h3>Gap three: predictability</h3><p>Not knowing when the next bus or shared ride will arrive costs more than its fare ever will. For most commuters who have another option, the anxiety of an open-ended wait is worse than paying for certainty. That is what a private car, or a hired driver, actually sells. Not speed. A known answer to “when will I get there.”</p><h3>Fix one: open the road, sized to real demand</h3><p>The fix for capacity is not a bigger single operator. A single public RTC already can’t fund and grow a fleet at today’s scale, and that is not a failure on its part, it is a shortfall of resourcing against demand that has simply outgrown it. The fix is letting more licensed operators run alongside it: buses and mini-vans sized to what a route actually carries, autos and shared cabs for the stretches those can’t reach, all sized to real demand rather than to a permit written years ago. Right of way to ply, the right to stop at shared bus stops, and simpler, cheaper registration for commercial vehicles are the concrete asks.</p><p>This is not a new model. Uber’s shared shuttle service already runs in Kolkata and Delhi, and was piloting in Hyderabad and Mumbai as of late 2024, with Bengaluru’s launch pending regulatory approval at that time. The model exists. What is missing in most cities is permission to run it at scale.</p><p>Opening the road this way has one real risk, and it is worth naming rather than skipping past. More operators on the kerb can mean more vehicles camping there, waiting to fill up, which adds to congestion instead of cutting it. The answer is to make the kerb itself automatic. A dwell allowance of 20 seconds, plus 10 seconds for each passenger boarding or alighting. Past that, a camera and an image classifier bill the vehicle continuously: ₹1 every 3 seconds it overstays. That works out to ₹20 a minute, the same rate as before, just metered by the second instead of rounded to the minute by a person with a notepad. Open access without that clock produces exactly the blocked kerbs critics predict. This is a proposed policy threshold. It is not a rule in force anywhere today.</p><h3>Fix two: smaller vehicles for the last stretch</h3><p>The last mile needs reach, not size. An auto, a shared taxi, or a small van can turn into a lane a 40-seat bus never will. It can also be added to a low-density neighbourhood that would never justify a full bus route on its own.</p><p>This does not compete with the bus or the metro. It completes the trip they start. Calling it a downgrade misreads the job. A trunk line and a last-mile shuttle are solving two different distances, and a city needs both.</p><h3>Fix three: a bus you can actually track</h3><p>The fix for predictability is the cheapest of the three. Live tracking, CrowdGatherGo’s own version is called WhatsComing and is still in development, maps each vehicle’s route, shares its live position, direction, speed, and current and next stop, and notifies a rider 2, 5, or 30 minutes before arrival, from home or the office.</p><p>What makes this practical is what it does not require. No new hardware fleet-wide. Only the standard smartphone a driver already carries. Commuters use the phones they already own. As a rough illustration of what that costs at scale: a city-sized fleet of about 7,000 buses, fitted at roughly ₹7,000 a smartphone, comes to about ₹5 crore. That is a planning-level estimate, not a vendor quote, and a small number next to the cost of new infrastructure.</p><h3>The food-delivery test</h3><p>Here is a test anyone can run on themselves. You track your dinner delivery to the minute. You do not track your bus.</p><p>The reason is not the food. The delivery app passes three tests the bus usually fails. There are enough riders out at the hour you order. The rider comes to your actual gate, not a depot down the road. And the app shows exactly where he is and when he will arrive. Take away any one of those three and most people would stop trusting the app, and go back to cooking.</p><p>A city bus fails the same three tests, for the same three reasons. Too few vehicles at the hour that matters. A stop that isn’t your gate. No way to know when it’s coming. Fix the same three things, capacity, reach, and a live screen, and the bus starts passing the test the delivery app already passes.</p><h3>What this would take</h3><p>Nothing here needs a new railway line or a wider road. It needs permission to run more licensed vehicles at the sizes demand actually calls for. It needs smaller vehicles doing the last-mile job a big bus was never built for. And it needs a phone screen that tells commuters the truth about when their ride is coming.</p><p>The day that combination is in place in even one Indian city, the real test of development gets easier to pass. Not whether a poor family can finally afford a car. Whether a family that already can, chooses not to.</p><p><em>CrowdGatherGo — Innovations in eXtreme Mobility. UNCLOG — DISPERSE — MOVE. </em><a href="https://crowdgathergo.com/"><em>crowdgathergo.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a971f1fbf9fb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Platform Can Be a Hole, Not a Hill: Level Boarding for Temporary Crowds]]></title>
            <link>https://medium.com/@atifhussain_54530/the-platform-can-be-a-hole-not-a-hill-level-boarding-for-temporary-crowds-4391589a90e1?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/4391589a90e1</guid>
            <category><![CDATA[crowd-management]]></category>
            <category><![CDATA[public-transit]]></category>
            <category><![CDATA[events]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Wed, 09 Sep 2026 02:00:41 GMT</pubDate>
            <atom:updated>2026-09-09T02:00:41.491Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3OJVlM8hEBq3xaA0nMFSYg.png" /></figure><h3>The problem with temporary millions</h3><p>A week-long crowd (Hajj, Kumbh, a match) never gets the infrastructure that would move it best. It gets queuing buses. One door, single file, dwell time stacking.</p><h3>What a train platform actually does</h3><p>A train platform is a raised edge that puts the coach floor and boarding surface at one height. Every door becomes a lane. A crowded train empties in seconds because the mechanism scales: more people means more coaches, each bringing more doors.</p><h3>Two ways to make the floor meet the ground</h3><p>Raise a platform (permanent corridors, permanent cost). Or sink a bay (temporary events, zero permanent infrastructure). The Road-Wagon’s 1,100 mm floor is the constant; the ground works are the variable.</p><h3>The Road-Wagon, in numbers</h3><p>7 doors, 92 passengers (32 seated + 60 standee), 12 m, battery-electric, level boarding. Door count — not seat count — sets how fast a crowd clears.</p><h3>The sunken bus-bay</h3><p>1,100 mm deep, ≥14 m per docked bus, ≤6% ramp, drained aggregate or concrete. Built in a day, filled back after the event.</p><h3>What this is not</h3><p>Not light rail; not a replacement for it. A frugal middle path for crowds that don’t justify permanent rail. Rail is better where the crowd is permanent; this wins where the crowd leaves.</p><h3>Where it fits</h3><p>Hajj, Kumbh and Arbaeen pilgrimages, inter-modal links, office commutes. Patent-pending (application 202641061415). Pre-homologation design specification.</p><p>— -</p><p><strong>Built by CrowdGatherGo — Innovations in eXtreme Mobility.</strong><br>UNCLOG · DISPERSE · MOVE</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4391589a90e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Seven Doors, One Rulebook: Certifying a Platform-Tied Bus in India]]></title>
            <link>https://medium.com/@atifhussain_54530/seven-doors-one-rulebook-certifying-a-platform-tied-bus-in-india-b77b8c6834f1?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/b77b8c6834f1</guid>
            <category><![CDATA[bus-design]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[public-transportation]]></category>
            <category><![CDATA[accessibility]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Mon, 07 Sep 2026 00:42:53 GMT</pubDate>
            <atom:updated>2026-09-07T00:42:53.963Z</atom:updated>
            <content:encoded><![CDATA[<p>We proposed hybrid of Bus and Metro Coach as solution to urban congestion, esp at demand points. Today we’ll look at it’s safety, convenience and regulation, to address whether it’ll be allowed in India.</p><h3>The question that comes before the engineering</h3><p>Every operator, body-builder and RTO asks the same thing first — will it be allowed? The question deserves a full answer, not an assumption. Here it is: seven doors is not an exception that requires ARAI to make a special case. It is an option that ARAI’s existing code already leaves open. The one number that genuinely departs from the norm is measured from a surface where passengers never stand.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ufGJ_U3TslwQKv3hZTfUiQ.png" /></figure><h3>What AIS-052 actually asks of a Type-I DLX bus</h3><p>The ARAI Code of Practice for Bus Body Design and Approval (AIS-052) requires a Type-I DLX standard bus to have a <strong>minimum</strong> of one service door, on the nearside. A second service door is <strong>optional</strong>.</p><p>This is not wordplay. A minimum is a floor. It is the lowest bar the code sets. Every vehicle that clears that bar and meets the other requirements (structural, emergency egress, accessibility) is code-compliant. Nothing in AIS-052 says “you can have only one door” or “two is the maximum.” The code does not say a maximum. It says minimum one, optional two.</p><p>The Road-Wagon exercises that option six times: D1 (the mandatory nearside stepped door) plus D2–D7 (six additional platform-level doors, geofenced and platform-presence-interlocked). Each door is oversized — platform doors are ≥1,200 mm wide (vs. 650 mm minimum), ≥1,950 mm high (vs. 1,800 mm minimum). They are not exceptions — they are compliant implementations of the space the code provides.</p><h3>The one genuine departure: floor height</h3><p>MoUD’s Urban Bus Specifications-II sets a cap on city-bus floor height: 900 mm. The Road-Wagon is 1,100 mm. This is a departure — from what?</p><p>The 900 mm cap was written for <strong>kerb-boarding</strong>. It assumes passengers board from the carriageway (street level). The Road-Wagon does not board from the carriageway. It boards from a platform at 1,100 mm (or, at a temporary event, from a bay sunk to the same depth).</p><p>Therefore: The vehicle is specified strictly under <strong>AIS-052 as a high-floor BRT-class bus</strong>, not under the MoUD kerb-boarding spec. It is compliant with the code it is written to, honest about the boundary, and transparent about why it matters.</p><h3>Measuring the step from the surface people actually stand on</h3><p>Floor height is measured from the carriageway: 1,100 mm — a “three steps from the street” figure on paper, true but describing a route nobody uses. The boarding surface is the platform (or sunken bay at temporary events). The actual step is ≤35 mm vertical, ≤75 mm horizontal. Less than the sole of most shoes. Both numbers are real because they describe different surfaces.</p><h3>Emergency egress and accessibility</h3><p>The Road-Wagon exceeds requirements: three dedicated emergency exits (E1, E2 rear windows; E3 roof hatch) plus mechanical release on all seven doors. Four priority seats, wheelchair bay at D2, level boarding at six doors. Accessibility is won by door count, not traded against it.</p><h3>Deployment before the platform is built</h3><p>D1 retains a conventional stepped entry with ramp. The vehicle works at an unimproved stop on day one, before any platform exists. The bus works before the corridor does.</p><h3>Two boarding interfaces, one floor height</h3><p>The vehicle meets a permanent raised platform (1,100 mm, ≥2,500 mm wide) at D2–D7, or a sunken bay for events (1,100 mm ±50 deep). The vehicle is unchanged; only the ground works differ. Platform-matching specification is precise; deployment is flexible.</p><h3>What we have not proved yet</h3><p>Pre-homologation design specification, May 2026. Dwell-time targets (≤18 seconds for the specified load) are design calculations, not field-measured. Simulation will verify; field trials will prove. ARAI pre-consultation will define the approval path.</p><h3>ARAI as evaluator</h3><p>ARAI is not a gatekeeper. ARAI is an <strong>evaluator</strong> asked to review the design against AIS-052 and approve the test of a pre-homologation prototype. Numbers on the table. Honest preconditions stated. Regulator asked to evaluate, not to gate-keep.</p><h3>Where this goes</h3><p>Prototype build, demo venues, high-traffic corridors, office commutes, inter-modal links, mass-gathering events (Hajj, Kumbh). Patent-pending (application 202641061415). Pre-homologation design specification complete. Ready for ARAI pre-consultation.</p><p>— -</p><p><strong>Built by CrowdGatherGo — Innovations in eXtreme Mobility.</strong><br>UNCLOG · DISPERSE · MOVE</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b77b8c6834f1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Timetable Is Written in Minutes. The Train Runs in Seconds.]]></title>
            <link>https://medium.com/@atifhussain_54530/the-timetable-is-written-in-minutes-the-train-runs-in-seconds-53e3f886fcb1?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/53e3f886fcb1</guid>
            <category><![CDATA[punctuality]]></category>
            <category><![CDATA[traffic-engineering]]></category>
            <category><![CDATA[indian-railways]]></category>
            <category><![CDATA[railways]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Sun, 06 Sep 2026 02:06:45 GMT</pubDate>
            <atom:updated>2026-09-06T02:06:45.308Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E2BzNdoz2mBayPrHQ2o7Bw.png" /><figcaption>In a packed schedule across a day, all delay errors cascade quickly. Even seconds of delay are significant. For Trains running at 60kmph with 5min headways, over a minute of delay reduces gap to under 3km changing green signal to double yellow.</figcaption></figure><h3>A forty-second train is an on-time train</h3><p>When a Working Time Table rounds arrival times to the nearest minute or half-minute, it is doing the best a paper schedule can do. The problem is that the train does not round. It runs in seconds. A locomotive that clears a block section 40 seconds late is “on time” by the timetable standard — and when your design headway is 300 seconds (5 minutes between slots), 40 seconds is not a rounding detail. It is 13% of the available gap.</p><p>That margin, repeated across a day’s traffic, adds up. Small errors cascade. A train 40 seconds behind compounds at the next stop. The follower behind it eats that delay, and the follower behind that eats both. One 40-second slip becomes network-wide lateness.</p><p>The problem is not the track. The problem is not the train. The problem is the precision of the data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TF4ZhQ87JU4jnu-R3zK8Ww.png" /></figure><h3>What a seconds-accurate km-post chart actually is</h3><p>Here is what it takes to build precision into a timetable without building new track.</p><p>First — conduct a non-stop survey run. Put the intended train (same weight, same driver, same consist) on the section with no stops. Measure the running time across each kilometre at the intended speed (say, 60 km/h on the main line, 30 km/h on the station loop). Interpolate the physics: if the run from km 0 to km 10 takes exactly 600 seconds, then each kilometre takes 60 seconds, and you know the train should clear km 214 at a certain time, to the second.</p><p>Second — emit a target. Instead of “arrival at Delhi platform: 06:40–06:41”, write “km 214 · 06:41:18”. The target is expressed at every km post, not just at stations. Every train in the schedule carries this precision.</p><p>Third — the driver reads this chart the way a ship’s captain reads a navigation chart. The targets are waypoints. The driver’s job is to pass each waypoint at the target time, or as close as the train’s physics allows. If he clears km 214 at 06:41:35 (17 seconds late), he knows it, and he can make up 10 of those 17 seconds over the next 10 kilometres if he needs to.</p><p>This is a data product, not a works programme. No new track. No capital cost. Just the information re-sharpened from minute-level to second-level.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BvPDXtyXFjCg7dsb8xrEHw.png" /><figcaption>Punctuality through Precision and real-time Feedback, unlocks hidden Capacity.</figcaption></figure><h3>The driver is the controller</h3><p>The chart is one half. The feedback is the other.</p><p>For decades, a driver receives updates at the station: “You are 2 minutes late” (learned after stopping). By then, the delay is baked in. The follower has inherited it. The opportunity to fix it live — while it is still small — is gone.</p><p>Instead: give the driver a live deviation display in the cab. Every 30 seconds (or at every km post), the display updates: “+22 seconds” or “−8 seconds”. The driver sees the number and corrects it himself, inside the existing safety buffer. A driver who sees “+22” eases off the throttle over the next few kilometres. A driver who sees “−8” increases speed within the safety envelope. Self-correction in real time, before the small error becomes a big detention.</p><p>The technology carrier for this feedback is WhatsComing — a RailTel physical broadcast that treats every track strictly one-way, so the loco always receives the latest signal without latency. No wireless dependency. No network bottleneck. The same broadcast carries the live signal aspect to the track circuits; it carries the live deviation to the cab.</p><h3>Both steps run on today’s Master Chart</h3><p>Here is what makes this powerful: it requires no new corridor. No new double line. No new loop infrastructure (though longer loops, where they exist, carry more advantage). The seconds-accurate chart and live driver feedback run on today’s Master Chart, with today’s train count, on today’s track.</p><p>A test band of ±60 seconds is what we are simulating (Q1–03, Kazipet–Wadi, SimPy microsimulation) to see how much variation a 300-second slot absorbs. That is a simulation band, not a measured result. But the mechanism is first-principles: if a driver sees live feedback and self-corrects, deviation does not compound the way it does when drivers learn about delays only at stations.</p><p>The deliverable is punctuality. A Working Time Table written to the second at every km post, with live driver feedback, is sharp enough to audit and honest enough to meet. It is what the Western Australian railways demonstrated at Surat–Vadodara, where the prerequisites (loop extensions, freight speed upgrades, automatic signalling) were applied and the result was tangible punctuality improvement.</p><h3>Then the throughput arithmetic</h3><p>Only after punctuality is proven does the throughput step become viable.</p><p>One uniform main-line speed (avoiding speed heterogeneity — mixing fast and slow trains). One 300-second slot (5 minutes, the design headway). Every halt absorbed on a ~1.5 km station loop (entered and run at 30 km/h so the following train’s gap is not cut). Four train classes (express, mail, passenger, freight, each with its own timing but all fitting the same slot rhythm).</p><p>A stop costs running time: about 5.6 minutes at 60 km/h, about 7.6 minutes at 120 km/h (this is braking and re-acceleration only, not the halt itself). Add the loop time (running 1.5 km at 30 km/h), and the follower’s gap is restored before the halted train rejoins the main line.</p><p>On the New Delhi–Chennai model (1,100 km, mixed traffic), this delivers roughly 500 train-slots a day on an existing double line — 250 per direction, with a 3-hour maintenance and contingency block. No new corridor. Same tarmac, different timetable.</p><h3>The preconditions</h3><p>This is not a fantasy. It presumes:</p><p><strong>Longer loops.</strong> A 1.5 km loop is longer than many existing loops. Where loops are shorter, the advantage shrinks. This is not a criticism — it is a physical constraint. It means the gain is highest on corridors with suitable loop geometry (many Indian main lines have it; some do not).</p><p><strong>Higher turnout speeds.</strong> Running into and out of a loop at 30 km/h requires smooth turnouts rated for that speed. Many existing turnouts are lower. Upgrading them is capital cost — less than building new lines, but real cost. It also requires RDSO evaluation of the tighter regime.</p><p><strong>Freight homogenisation and windowing.</strong> Mixing fast and slow freight is what destroys slots. The solution is freight windowing — dedicated time windows where freight trains run together, not interleaved with passenger. This is an operational change, not a track change, but it is structural.</p><p><strong>Heavy-haul coal as the residual limit.</strong> On corridors where coal is 40% of goods traffic (like Kazipet–Balharshah), coal trains at max load set the speed. The slot timetable has to accommodate them. This is not a flaw — it is what the numbers constrain. It means throughput gain is most visible on corridors where coal is not the sole cargo.</p><h3>Where seconds-level feedback pays first</h3><p>The honest truth is that punctuality delivers first, throughput comes later. A timetable written to the second, with live driver feedback, pays off immediately on any line. The throughput gain requires the additional stack of longer loops, turnout speeds, freight windows, and slot discipline. But punctuality? That is a data win.</p><p>A division running on a traditional WTT might see 15–20% of trains latterly recorded as “on time” (within the loose minute-level standard). Switch to seconds-accurate charts and live feedback, and that number can jump to 70–80%, with no new construction.</p><p>That is why this is the first layer of the fix. Not because throughput is unimportant. But because it is honest, it is measurable, and it works on today’s track.</p><h3>RDSO as evaluator</h3><p>The regulatory path is clear: RDSO is not a gatekeeper here. It is an evaluator. The question is not “Will you let us do this?” The question is “We want to test this tighter regime on this section. What does RDSO need to evaluate to approve the test?”</p><p>The answers are specific: What loop geometry is needed? (Design specification, Clause 17, Road-Wagon example) What turnout speed is required? (Civil works specification, feedback from RDSO field teams) What freight windowing regime? (Operations plan, coordination with goods depots) What is the test band? (±60 seconds on a 300-second slot; simulation results from Kazipet–Wadi)</p><p>Bring these numbers to RDSO as a pre-consultation package. Ask them to evaluate. Accept their feedback and iterate. This is the posture that works.</p><h3>Why this matters</h3><p>Punctuality is not a comfort issue. It is a network capacity issue. When trains run late, the network cannot pack traffic tightly because of the uncertainty. Dispatchers have to leave larger gaps to avoid conflicts. That larger gap means fewer trains in the same corridor.</p><p>Fix punctuality, and the corridor inherently carries more traffic, without new track. The same double line now carries 500 slots instead of 250 because the slots can be 300 seconds apart instead of 600, and the tighter spacing is stable because the timetable is precise and drivers correct in real time.</p><p>This is frugal infrastructure: better data, better feedback, better operations. The same metal, better use.</p><p>— -</p><p><strong>Status:</strong> Testing phase Q1–03 (SimPy microsimulation, Kazipet–Wadi, ±60 second test band). Pre-consultation with RDSO to begin Q4–01. Field trial on suitable corridor target: Q2–02.</p><p><strong>Built by CrowdGatherGo — Innovations in eXtreme Mobility.</strong><br>UNCLOG · DISPERSE · MOVE</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=53e3f886fcb1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Seven Doors: Why the Bottleneck at the Stop is Not the Bus]]></title>
            <link>https://medium.com/@atifhussain_54530/seven-doors-why-the-bottleneck-at-the-stop-is-not-the-bus-bbc4ba4f34af?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/bbc4ba4f34af</guid>
            <category><![CDATA[urban-planning]]></category>
            <category><![CDATA[mobility]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[urban-mobility]]></category>
            <category><![CDATA[public-transportation]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Sat, 05 Sep 2026 02:36:34 GMT</pubDate>
            <atom:updated>2026-09-05T02:36:34.221Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-usdSugjYc1BrkXJg4ZjpQ.png" /></figure><h3>The stadium gate problem</h3><p>Watch a crowd file out of a stadium at the end of the day. They move through one gate, get stuck, back up, wait. The gate is not bad at letting people through — it is the only gate doing the job. One opening for thousands.</p><p>Now watch a train platform empty in the same time: every door opens at once, passengers stream through seven exits simultaneously, the platform clears in seconds. Same volume. Different architecture.</p><p>The difference between a bottleneck and a flow is not how fast the door opens. It is how many doors there are.</p><h3>Why we blame the bus instead of the door</h3><p>In cities, we count vehicles: buses per route, route frequency, capacity per bus. The 50+ people a bus carries in the road space of four cars makes it the single most efficient vehicle on a street. By every vehicle metric, a bus wins.</p><p>Then it stops at a kerb, and the whole lane stops with it.</p><p>Dwell time — the seconds a bus sits at a stop while passengers board and alight — is treated as an operational efficiency problem: better door design, faster boarding, better crowding management. We shave seconds off one door’s throughput.</p><p>But one door has a ceiling. A single 800 mm opening, even perfectly efficient, can only move so many people per minute. A stadium gate does not become a good exit just because the people move faster.</p><h3>What a train platform actually solves</h3><p>A train platform is not primarily about height — though that matters. It is about <strong>doors</strong>. A coach floor and platform are at the same level so every door can operate as a full boarding lane. A crowded train empties in seconds not because trains are magically efficient, but because the mechanism scales with volume: more people means the train adds more coaches, each bringing more doors.</p><p>50 people leave through one door (bus): they queue, they wait, the vehicle is immobilised. 500 people leave through five doors (small train): they walk through simultaneously, the platform clears in the time it takes to board.</p><p>Dwell time is not a boarding speed problem. It is a door count problem.</p><h3>The 7door-bus-&amp;-platform</h3><p>The mechanism is simple: put seven doors on a bus, build a platform to the same height as the bus floor, and the boarding physics become a train’s.</p><p><strong>Seven doors</strong> — each wide (≥1,200 mm, wider than the code minimum of 650 mm) and tall enough for easy passage (≥1,950 mm against 1,800 mm minimum). This is AIS-052 compliance, not an exception. The code requires one service door and makes a second optional; six additional doors are an exercise of that option, within specification.</p><p><strong>Level platform</strong> — 1,100 mm high, the same height as the bus floor. No steps to trip on, no bend at the door. Passengers walk straight through from platform to bus floor, as they do on a train. The boarding interface is flat.</p><p><strong>Four-door pairs in layout</strong> — the seven doors serve four main boarding lanes (D2–D7), each with its own approach path and crowd-management space, plus a stepped entry at D1 for conventional stops where platforms don’t yet exist.</p><p><strong>Result:</strong> 92 passengers (32 seated, 60 standing) load and unload simultaneously through four boarding lanes. Dwell time collapses from ~60 seconds (one door, one queue) to approximately 18 seconds (four door pairs, four crowds moving at once). The bus clears the next junction before the signal turns.</p><h3>The platform, once</h3><p>The platform is a kerb-side structure — ≥2,500 mm wide, ≥18 metres of docking space per bus — at 1,100 mm height. It is built once and serves every 7door-bus that corridors use. The vehicle carries the platform-matching specification; the infrastructure carries the height match.</p><p>For routes where platforms do not yet exist, D1 retains a conventional step well (≤400 mm first rise, ≤250 mm intermediate, kneeling and ramp for accessibility). The vehicle works on improved stops and unimproved stops. This is not a trade — it is a deployment choice.</p><p><strong>No overhead wires.</strong> The vehicle is 12 metres, battery-electric. Charging is depot-based or en-route inductive. The corridor does not carry infrastructure costs.</p><p><strong>No new corridor.</strong> The bus fits standard road width. The platform replaces a conventional bus shelter — it is a raised kerb, not a new structure.</p><h3>Dwell time is the move</h3><p>Dwell time is where door count becomes a corridor mechanic. When a bus is stuck at a stop, the block section behind it is stuck. When five buses queue at one stop because each one takes a minute to clear, every bus behind them inherits that delay.</p><p>Reduce dwell time to 18 seconds, and a bus can serve a stop and move to the next one in the time a conventional bus is still loading. Capacity per vehicle jumps not because the bus is faster, but because it loses less time sitting still.</p><p>In a city where congestion is a stopping problem — where vehicles halt inside the flow lane and everyone behind them pays the cost — door count is not a vehicle feature. It is an infrastructure move.</p><h3>One window of proof</h3><p>This is a pre-homologation design specification, filed for patent (application 202641061415, dated 14–05–2026). The dwell-time figure (≤18 seconds for the specified load, four door pairs, 30 passengers per door pair per cycle) is a design target based on the door geometry and boarding rate engineering. It is not yet a field-measured result.</p><p>Turning it into a result is the next step: build a prototype, test it on a demonstration corridor, and measure what the platform and seven doors actually deliver on real crowds and real stops.</p><p>The mechanism — more doors, faster clearing — is first-principles physics. The specific numbers are engineering targets that need verification in the field.</p><h3>Why this matters</h3><p>A city’s congestion is not a shortage of vehicles. It is vehicles stopped in the wrong place for the wrong reason. A bus stop that forces 92 people through one 800 mm door is not a boarding problem. It is an architecture problem.</p><p>Solve it at the door, and the stop becomes a flow. Solve it on every major route, and the network changes.</p><p>This is frugal infrastructure: one platform design, one vehicle specification, and no new corridor or overhead cost. The same tarmac, better use.</p><p>— -</p><p><strong>Status:</strong> Pre-homologation design specification. Patent-pending (application 202641061415). No ARAI approval has been granted. Field trials to follow.</p><p><strong>Built by CrowdGatherGo — Innovations in eXtreme Mobility.</strong><br>UNCLOG · DISPERSE · MOVE</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bbc4ba4f34af" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[One Number, Two Different Roads]]></title>
            <link>https://medium.com/@atifhussain_54530/one-number-two-different-roads-4666b25da3c1?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/4666b25da3c1</guid>
            <category><![CDATA[traffic-jam-calculations]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[traffic-stream-review]]></category>
            <category><![CDATA[traffic-engineering]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Fri, 04 Sep 2026 05:39:19 GMT</pubDate>
            <atom:updated>2026-09-04T05:55:03.462Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/942/1*zntChTUQVThC1KAzXC5EUQ.png" /><figcaption>Traffic can’t be modelled with just one number. It can’t tell a working junction from a failing one.</figcaption></figure><h3>Why a traffic count cannot tell a working junction from a failing one.</h3><p>Somewhere in your city there is a junction that flows at midnight and locks by nine in the morning.</p><p>If you had to prove that, you would probably reach for a traffic count. Vehicles per hour. It is the standard unit of traffic evidence. It is cheap to collect, easy to file, and it looks solid.</p><p>It is also the one number that cannot answer the question.</p><h3>The number every city collects</h3><p>A counter at the kerb records how many vehicles pass in an hour. That figure then travels a long way. It goes into surveys, into proposals, into decisions about which road to widen.</p><p>The trouble is not that the count is inaccurate. The trouble is that the same count describes two opposite roads.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XclKsvRmwJOJKtLj6nhLWw.png" /></figure><h3>Flow is two numbers multiplied</h3><p>Three things describe a stream of traffic.</p><p><strong>Volume</strong>, or flow. How many vehicles pass a point in an hour. Call it q.</p><p><strong>Density.</strong> How tightly packed they are. How many sit on one kilometre of one lane at this instant. Call it k.</p><p><strong>Speed.</strong> How fast the stream is actually moving. Call it v.</p><p>They are tied together by one line: <strong>flow equals density times speed.</strong> q = k × v.</p><p>Read that again, because it is the whole article. Volume is a product. It is two numbers multiplied. The moment you record only the product, you have thrown away both the numbers that made it.</p><p>Ten times three is thirty. So is two times fifteen. Hand someone the number thirty and they cannot tell you which one you started from.</p><p>That is what a vehicle count does to a road.</p><h3>The same count, twice</h3><p>Now put numbers on it.</p><p>Take the simplest traffic-stream model there is, Greenshields. Speed falls in a straight line as density rises. An empty road runs at its free speed. A jammed road stops. In between, it is linear.</p><p>Set the free speed at 60 km/h and jam density at 200 vehicles per km per lane. Those two values are illustrative. They are round numbers chosen so the arithmetic is easy to follow, not measurements of anybody’s road. The point that follows does not depend on them.</p><p><strong>Case one.</strong> Fifty vehicles per km. The road is a quarter full, so the stream runs at 45 km/h.</p><p>50 × 45 = <strong>2,250 vehicles an hour.</strong></p><p><strong>Case two.</strong> A hundred and fifty vehicles per km. Three-quarters full. The stream has slowed to 15 km/h.</p><p>150 × 15 = <strong>2,250 vehicles an hour.</strong></p><p>The same figure.</p><p>One of those roads is working. The other is crawling at 15 km/h, three-quarters packed, and one small obstruction away from stopping. The counter at the kerb writes the same number for both and files it.</p><p>This is not a quirk of the chosen values. Below its peak, a road produces every flow figure twice. Once on the way up, once on the way down.</p><h3>The best a road ever does is half its free speed</h3><p>Work out where the peak sits and you get a clean answer. Flow is highest when density is exactly half of jam density. And at half of jam density, the straight-line rule puts speed at exactly half the free speed.</p><p>In our example: 100 vehicles per km, moving at 30 km/h, carrying 3,000 vehicles an hour. That is the most that lane will ever do.</p><p>Sit with that. On a 60 km/h road, the moment of maximum throughput is the moment traffic is moving at 30.</p><p>So a stream at half its free speed is not failing. It is the road working as hard as it can.</p><p>The alarm is the speed <strong>below</strong> half. Past the peak the curve turns over, and every extra vehicle now subtracts capacity instead of adding it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a9VObaqbTQM7T1uKUAt9bg.png" /></figure><h3>What tips a road over</h3><p>This is also the mechanism behind something recorded on the KRPuram to TinFactory stretch in Bengaluru in May 2025. A small obstruction, and stream speed fell from 60 kph to 15 kph.</p><p>Nothing was added. No extra traffic arrived. One near-stopped vehicle made the traffic behind it bunch up, and bunching is rising density. Density crossed the peak, speed collapsed, and the road slid down the far side of the curve on its own.</p><p>It is the same thing every junction does. A four-way signal stops three arms so one can move, and gives each arm about a quarter of the time. A single unsignalled U-turn puts one near-stopped vehicle across a fast stream. Different designs, same physics: a slow body in a quick stream, density pushed past the peak.</p><h3>Where the model stops and the road starts</h3><p>Now the honest part.</p><p>Greenshields is a straight-line approximation. It assumes a tidy relationship between speed and density. Indian traffic is not tidy. Two-wheelers use gaps a car cannot. Vehicle lengths vary by a factor of five. Lane discipline is often notional. The real curve for your road has a different shape, a different peak and a different jam density from the round numbers above.</p><p>That is not an argument against the method. It is the argument for it.</p><p>If the textbook curve does not describe your junction, you certainly cannot run your junction on a textbook curve. And you cannot run it on a count either. You have to measure the road you actually have.</p><p>Two things survive whatever shape the curve takes. There is a peak. And below the peak, every flow figure happens twice.</p><h3>What a camera can already see</h3><p>So how do you measure it, without new gantries, new sensors or new civil work?</p><p>You point a camera at the road. On most busy junctions, someone already has.</p><p><strong>LanePulse</strong> is our traffic-measurement software. It takes camera input, identifies vehicles and their direction and speed, and reports road width, average speed, vehicle density and traffic volume at each point on the road.</p><p>It is <strong>in development</strong>. It is not running anywhere. It needs approved camera feeds to prove itself on, and getting that approval in writing is the next step on our own plan, not something behind us. There is no accuracy figure here, because none has been earned yet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LUrgpawMensYBhtKpH-nNg.png" /></figure><h3>From an opinion to a before and after</h3><p>Here is why the measurement is worth the trouble.</p><p>Once you can read density and speed, the conversation about a junction changes shape. Instead of “this one is terrible in the mornings”, you get: at 9.15 this arm sits at this density and this speed, which puts it here, on the falling side of its peak.</p><p>Then you change something. Cut four signal phases to two. Replace a single U-turn with a signalled, parallel, size-graded one. Build an interchange where the geometry really needs it.</p><p>Then you measure the same three numbers again.</p><p>A redesign stops being an opinion. It becomes a number that moved.</p><p>That matters beyond one junction. Change carries risk, and without a measurement nobody can show that the last change worked. So the second junction is as hard to approve as the first, forever. A number that moves is what turns one good junction into a programme.</p><h3>What we are asking for</h3><p>The instinct is to widen the road. The cheaper first step is finding out which side of the peak it is already on.</p><p>It is the same discipline we use on the railway side, where a timetable is written to the second at every kilometre post and the deviation is read back live. Measure the thing you intend to change. Change the least you can. Measure again.</p><p>The aim is a city where average traffic speed recovers towards 40 kmph. Not because anything was demolished, but because somebody measured first.</p><p>So: which junction in your city flows at midnight and locks by nine? And which camera is already pointed at it?</p><p><em>CrowdGatherGo — Innovations in eXtreme Mobility. UNCLOG — DISPERSE — MOVE. </em><a href="https://crowdgathergo.com/"><em>crowdgathergo.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4666b25da3c1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[One Person, Every Department Run by AI]]></title>
            <link>https://medium.com/@atifhussain_54530/one-person-every-department-run-by-ai-a3508cb287ed?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/a3508cb287ed</guid>
            <category><![CDATA[solopreneur]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[bootstrapped]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[b2g]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Thu, 03 Sep 2026 06:26:24 GMT</pubDate>
            <atom:updated>2026-09-03T06:26:24.459Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*j289-T7DZV8T8boB-hmSlw.png" /></figure><h3>Why a mobility company selling to government is built the way it is.</h3><p>Operations, patent drafting, event preparation, marketing content, and the feasibility challenge to my own ideas. GenAI does the work, mostly Claude. I supply the domain knowledge and the judgement.</p><p>People usually hear that as a cost story. It is not. It is a distance story.</p><h3>The road you cannot shorten</h3><p>CrowdGatherGo sells to government. Traffic departments, railway zones, event authorities.</p><p>A public buyer is spending public money, so it checks, and the checking takes time. A file moves through desks. A pilot is scoped, then scoped again. That care is correct and I would not want it removed. It is also, from the seller’s side, a fixed length of road.</p><p>You cannot shorten it. No amount of engineering quality shortens it. The idea can be right, the drawings can be right, and the gap between the first meeting and the first rupee stays exactly as long as it was.</p><h3>What actually kills infrastructure ideas</h3><p>Not the engineering.</p><p>An idea in this field usually dies because the firm carrying it ran out of cash before the buyer finished deciding. The work was sound and the tank was empty. That has ended more infrastructure proposals than bad physics ever has.</p><p>So the question for a bootstrapped firm is not “is the idea good”. It is “does the fuel last the distance”.</p><h3>Fuel per kilometre</h3><p>Here the arithmetic changed.</p><p>Work that needed a team of ten now needs one person and a good prompt. The patent draft, the deck, the feasibility check, the outreach, the content. Each of those used to be a salary. Now each is a session.</p><p>That is a cut in fuel burned per kilometre. It is not a shorter road. AI does not make a government department decide faster, and anyone who tells you it does has not sold to one. What it does is let the tank outlast the journey.</p><p>The distinction matters, because it tells you what to spend the saving on. Not on going faster. On being able to still be there when the answer comes.</p><h3>The floor has dropped, and not only for me</h3><p>This is not a personal quirk. The minimum size of a serious firm has fallen, and it has fallen fastest where the hard part is thinking rather than manufacturing.</p><p>James Currier at NFX called it the three-person unicorn. With AI running each function, a very small team can carry a company that used to need hundreds. He also points out that the pattern is older than AI. Instagram had 12 employees when it sold for $1 billion. WhatsApp had 50 when it sold for $21 billion. What AI changed is how many kinds of business this is now open to.</p><p><strong>Medvi</strong>, a telehealth firm, was started in September 2024 with $20,000 and no outside money. It has two employees, the founder and his brother. It did $401 million of revenue in its first full year at about 16 percent net margin, and is on track for $1.8 billion this year. Bootstrapped, high growth, high margin, two people.</p><p><strong>Cursor</strong> passed $100 million of annual recurring revenue with 30 to 50 people in early 2025. It is far larger now, and that is the point. It reached the scale before it needed the headcount.</p><p><strong>Safe Superintelligence</strong> had around 20 people when it raised at a $32 billion valuation in March 2025.</p><p>The tail is just as telling as the outliers. China had over 16 million one-person companies registered by June 2025, more than a quarter of all businesses there, and added 2.86 million more in the first half of 2026, up 47 percent on the year before.</p><p>Currier’s three-person team is a visionary generalist, a words person and a numbers person. I am the generalist. AI is the other two, and I check its physics.</p><h3>What this buys a mobility company</h3><p>Everything CrowdGatherGo works on is a long-cycle, high-scrutiny sale. Slot-based timetabling on an existing double line. Junction redesign measured before and after. A patent-pending 7door-bus and platform system for boarding without blocking the lane.</p><p>None of those get bought in a quarter. All of them need the firm to still exist in three years, with the drawings maintained, the simulation runnable, and the founder still across the detail.</p><p>One person plus AI is what makes that survivable on a bootstrap. It is the operating decision that lets the engineering decisions get a hearing.</p><h3>What I still do myself</h3><p>The judgement, and the domain knowledge that judgement runs on.</p><p>Whether a headway is safe. Whether a loop is long enough. Whether a claim is stronger than the evidence behind it. Whether a diagram matches the mechanism it is drawing. AI is fast and it is confident, and confident is the dangerous part. Every technical claim that leaves this company is checked by a human who knows the field, because a wrong figure in front of a railway officer costs more than the whole content programme is worth.</p><p>The tools run the departments. They do not sign off the physics.</p><h3>If you are in the same position</h3><p>If you run a small firm with big, slow customers, I would like to hear how you are handling the wait. Not the pitch. The wait.</p><p>That is where these things are actually won and lost.</p><p><em>CrowdGatherGo — Innovations in eXtreme Mobility. UNCLOG — DISPERSE — MOVE. </em><a href="https://crowdgathergo.com/"><em>crowdgathergo.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a3508cb287ed" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[BUS: The Best Vehicle on the Road, and the Worst Stop]]></title>
            <link>https://medium.com/@atifhussain_54530/bus-the-best-vehicle-on-the-road-and-the-worst-stop-239967dafdaf?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/239967dafdaf</guid>
            <category><![CDATA[public-transport]]></category>
            <category><![CDATA[urban-mobility]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[bus-stop]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Wed, 02 Sep 2026 04:57:41 GMT</pubDate>
            <atom:updated>2026-09-02T04:57:41.172Z</atom:updated>
            <content:encoded><![CDATA[<h3>The Best Vehicle on the Road, and the Worst Stop</h3><h3>Boarding time is governed by doors, not by buses.</h3><p>Two statements, both true, and the second one is why the first is wasted.</p><p>A bus is the most space-efficient vehicle on any Indian street. Fifty or more people in the road space of about four cars. If you were designing an urban road network from scratch, you would build it around buses and it would not be close.</p><p>A bus stop is the worst piece of infrastructure on the same street.</p><h3>What happens when it stops</h3><p>When a bus pulls in at the kerb, it stops the lane.</p><p>Not for a moment. For as long as boarding and alighting take. And that is a long time, because everyone gets on through one door, or two. A queue through a single door is serial — one person at a time, and every person who fumbles for change or shifts a bag holds up everyone behind them.</p><p>So the vehicle that uses road space better than anything else spends a large part of its day occupying a full lane and moving nobody.</p><p>Now scale it. Ten buses at a busy stop at peak, which is exactly what happens on a corridor that is working well. All ten are stopped. Everything behind them is stopped. The better the service, the worse the blockage.</p><h3>Why a dedicated terminal moves the problem</h3><p>The standard escalation is understandable and it is what most cities do. Raise frequency, add buses, and the stop grows into a dedicated terminal or bus bay complex off the main road.</p><p>That does help the through lane. It also buys land in the most expensive part of the city, lengthens the walk at both ends of the trip, and gathers all the stopping into one place. The congestion moves. It does not leave.</p><p>And notice that none of this touches the thing that caused it. The bus still boards through one door. The terminal has simply given it somewhere less damaging to do that.</p><h3>The arithmetic of doors</h3><p>Here is the part that reframes the problem, and the useful thing about it is that it has nothing to do with buses.</p><p>Boarding time is governed by the number of parallel channels, not by the size of the vehicle.</p><p>If a vehicle has one door, boarding is one queue and the time grows with the number of passengers. If it has seven, you have seven queues running at once, and the same crowd clears in roughly a seventh of the time, minus whatever the slowest door costs you.</p><p>That is not a claim about a product. It is queueing arithmetic, and you can watch it working at scale every day.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Og_i4pADTJLPi0WU.png" /></figure><h3>The airliner and the metro coach</h3><p>Take two vehicles that carry comparable numbers of people.</p><p>An airliner boards through one jet bridge into one aisle. It takes something like half an hour, and everyone has experienced why: one person stops to lift a bag into the locker, and the entire aisle behind them waits.</p><p>A metro coach has four doors on each side, opening on to a level platform with no step and effectively no gap. The same order of passenger count clears in seconds.</p><p>Nothing about the passengers is different. The task is the same task. The difference is doors, and the absence of a step.</p><p>Our city buses are built like the airliner.</p><h3>Level boarding, and why the step matters</h3><p>Doors are most of the story, but not all of it.</p><p>A step costs time on every single passenger, and it costs far more than average time on the passengers who are slowest — an older person, someone carrying a child, someone with a stick, anyone with luggage. Those passengers set the door’s throughput, because the queue behind them cannot overtake.</p><p>A vehicle tied to a platform, boarding level and without a gap, removes that from every door at once. This is the second thing the metro coach is doing that the bus is not, and it compounds with the first.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*j8gPY2n6Qq7Chdbk.png" /></figure><h3>Where this matters most</h3><p>The everyday case is the office commute and the interchange — a metro station, a railway station, an airport, where a large number of people arrive together and need to be moved a short distance.</p><p>The extreme case is a mass gathering. Kumbh. Hajj and its rituals. Arbaeen. Stadium events. These are the conditions the whole idea was shaped around, because at that scale dwell time is not a comfort question. It is the thing that decides whether a crowd disperses or compresses.</p><p>At an ordinary stop, a slow-boarding bus costs minutes. At a gathering of millions, it decides how a crowd behaves.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nkWTGrJ0wzqA7u_M.png" /></figure><h3>What is built, and what is not yet proven</h3><p>So, plainly, because this is the part where mobility writing usually overreaches.</p><p>The Road-Wagon is a platform-tied multi-door road vehicle — seven doors and a matched platform, so boarding runs in parallel and level, and the vehicle leaves the lane sooner. In layman terms, the 7door-bus.</p><p>It is patent pending. Indian patent application 202641061415, filed 14 May 2026.</p><p>It is in development. No approval has been granted. And I am publishing no dwell-time figures for it, because there is not yet a vehicle to measure. The queueing arithmetic above stands on its own, and the metro coach proves it daily. What our specific vehicle achieves is a measurement I have not taken yet, and I will not claim it before I have.</p><p>What I will say is which question is worth arguing about. Not whether buses are good — they are, on the numbers, better than anything else on the street. The question is why the best vehicle on the road is still built to load like an aeroplane.</p><p><em>CrowdGatherGo — Innovations in eXtreme Mobility. UNCLOG — DISPERSE — MOVE. </em><a href="https://crowdgathergo.com/"><em>crowdgathergo.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=239967dafdaf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Unlocking Urban Mobility: Two Halves of One Policy]]></title>
            <link>https://medium.com/@atifhussain_54530/unlocking-urban-mobility-two-halves-of-one-policy-5f5ab36ee959?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/5f5ab36ee959</guid>
            <category><![CDATA[public-policy]]></category>
            <category><![CDATA[urban-mobility]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <category><![CDATA[public-transport]]></category>
            <category><![CDATA[smart-cities]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Tue, 01 Sep 2026 05:00:28 GMT</pubDate>
            <atom:updated>2026-09-01T05:00:28.086Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BYAx699QLYfu4FH580dHDQ.png" /></figure><h3>Open the kerb to anyone who will run a bus. Then put a clock on it. Neither half works alone.</h3><p>India retailed 25,91,138 vehicles in July 2026. That is up 25.89% on the same month last year, about 83,600 a day, and a best-ever July in all six categories.</p><p>It is a striking number, and it is easy to read wrong. So before anything else, here is what it does not prove.</p><h3>What the number does not prove</h3><p>It does not prove that anybody distrusts their city’s transport network. Registrations cannot carry that weight, and claiming they can is the fastest way to lose a serious reader.</p><p>FADA’s own attribution should be conceded first. The surge is credited to GST-led affordability, to rural sentiment, and to the festive build-up. On top of that, vehicle demand is income-elastic — it rises faster than income as a matter of course, and a multiple above one is entirely normal. India’s real GDP grew 7.7% in FY26 while vehicle retail grew roughly three and a half times as fast. That gap is expected, not scandalous.</p><p>And one line has to be said plainly, because it is the line that keeps this argument honest: <strong>nobody buying a car is voting against a metro.</strong> Corridors under construction are real, they are needed, and the people building them are doing hard work well.</p><p>What the number <em>is</em> good for is something narrower and more useful.</p><h3>Pool and flow are different quantities</h3><p>A city’s vehicle <strong>pool</strong> barely moves when its network improves. A car that has already been bought is sunk, financed, insured and parked. A new corridor opening does not send it to the resale market. The pool changes over years.</p><p>The <strong>flow</strong> of new registrations is a different animal. It responds, in either direction, and it is recorded monthly, city by city, at national coverage, funded by the people making the decision. No survey has those properties.</p><p>The pool is the balance sheet. New registrations are the share price.</p><p>One caveat on that comparison, because it has a real limit. A share price moves on the tick, because the market is continuously priced. Vehicle purchases are lumpy, financed and seasonal, and registrations trail the decision through booking, finance and delivery. So claim the reactivity, never a same-month resolution.</p><h3>Cars against cars</h3><p>Comparisons in this area go wrong when the units are mixed, so here is one that does not.</p><p>Bengaluru has about 25.2 lakh private cars as of 30 June 2026, the largest car fleet of any Indian city. National passenger-vehicle retail in July 2026 was 4,16,555.</p><p>So Bengaluru’s entire private car fleet is about six months of national car sales. Cars against cars, one country against one city. Six months of India buys one of our largest cities over again.</p><p>That is the scale the shared-transport answer has to be sized against.</p><h3>Why buses, and not only corridors</h3><p>Metros and suburban rail matter, and this is not an argument against building them. It is an argument about response time.</p><p>A corridor is supply measured in kilometres. A commute is demand measured in doors — your door to their door, at the hour you actually travel. Both can be true at once: the corridor excellent, and the trip still failing at the first mile, the last mile, the interchange, or the evening frequency.</p><p>Buses and smaller shared vehicles — mini-vans, autos, bike-taxis — are the fastest way to change what a trip feels like, because they can be added in months rather than years, and sized to whatever the market actually wants.</p><p>The constraint is not willingness. It is that demand at today’s scale has outgrown what any single operator can fund and grow. Delhi’s public bus fleet, once above 6,000, now runs in the region of 4,000. Mumbai’s BEST fleet is below 2,600. Those are capacity gaps against a city that has doubled around them, not anybody’s failure.</p><p>India met exactly this shape of problem in 1991, in the economy at large. The answer then was to let private capital pool and allocate. The same move is now due in urban shared transport.</p><h3>Half one: right of way</h3><p>The ask is specific and it is small.</p><p><strong>Right to ply.</strong> Private operators may run shared transport — buses, mini-vans, autos, bike-taxis — at whatever size the market demands, rather than at a size fixed in advance by a permit category.</p><p><strong>Right to stop.</strong> They may use shared bus stops, not separate ones. A passenger should not have to know which company runs a vehicle in order to know where to wait.</p><p><strong>Cheaper entry.</strong> Simplified and reduced registration and ownership costs for yellow-board commercial vehicles, so a two-vehicle operator is a viable business.</p><p>Stated on its own, that ask has a well-known objection, and the objection is correct.</p><h3>Half two: a clock on the kerb</h3><p>Everybody has watched it happen. A shared vehicle pulls up at a stop and simply stays, waiting to fill, while the road behind it does not move. Rented scooters get abandoned on the footpath. Open the market without addressing this and you get exactly that, multiplied.</p><p>The answer is not more penalties. It is that this is now measurable, cheaply, which it was not when the permit system was designed.</p><p>Cameras at bus stops, or on a nearby high-rise, running image classifiers, can enforce a simple rule:</p><p><strong>A dwell allowance of 20 seconds, plus 10 seconds for each person boarding or alighting.</strong> Past that, the system sounds a warning, then charges at ₹20 per minute of continued occupancy.</p><p>For shared scooters left in the wrong place: booked on the spot, licence suspended, and the user removed from all platforms at once.</p><p>Notice what the rule governs. Not who may stop. How long they may stay.</p><h3>Why neither half works alone</h3><p>This is the whole point, and it is why these two are one policy rather than two.</p><p><strong>Open access without a clock</strong> produces the blocked kerb the objection predicts. Every operator has an incentive to wait for one more passenger, and the road pays for it.</p><p><strong>A clock without open access</strong> changes nothing at all. You have carefully regulated the dwell time of a fleet that is already too small.</p><p>Put them together and the kerb becomes a high-throughput piece of infrastructure rather than a parking space. The same twenty metres serves many operators, all day, because nobody is allowed to camp on it.</p><h3>The hospital porch</h3><p>If you want the shape of it in one picture, look at the drop-off porch outside any large hospital.</p><p>It is a short stretch of kerb. Hundreds of vehicles use it in a day — ambulances, taxis, autos, family cars. Everyone is allowed to pull in.</p><p>The only rule that matters is how long you may stay. Take that rule away and the porch fills with parked cars waiting for someone to come out, and the ambulance cannot reach the door. Same kerb, serving almost nobody.</p><p>Open access plus a clock. That is the entire design.</p><h3>A testable prediction, and a ratio worth publishing</h3><p>Here is the claim, in falsifiable form. On the day such a policy takes effect, operators enter the new market, <strong>new bookings for private white-board vehicles turn first, and registrations follow within a quarter.</strong> Registrations lag through booking, finance and delivery, so a same-month drop is not what to look for.</p><p>And here is a smaller ask that costs nothing and could start on Monday.</p><p><strong>Publish one ratio, monthly: private-vehicle registration growth divided by the city’s own income growth.</strong> Both numbers already exist. Nothing new is collected. A network winning trust should push that ratio down, and it should move before ridership does — a leading indicator, in a field that has almost none.</p><p>If it rises after a corridor opens, that is not a verdict on the corridor. It says the corridor has not yet reached the trips people are actually making.</p><p>One last thing, since it comes up every time. Electrification is a decarbonisation win, and a real one. It is not a congestion win. A clean car occupies exactly the same lane.</p><p><em>CrowdGatherGo — Innovations in eXtreme Mobility. UNCLOG — DISPERSE — MOVE. </em><a href="https://crowdgathergo.com/"><em>crowdgathergo.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5f5ab36ee959" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Slot Belongs to the Track, Not the Train]]></title>
            <link>https://medium.com/@atifhussain_54530/the-slot-belongs-to-the-track-not-the-train-411ca9837fc3?source=rss-8c9f7d2b55fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/411ca9837fc3</guid>
            <category><![CDATA[railway-engineering]]></category>
            <category><![CDATA[indian-railways]]></category>
            <category><![CDATA[infrastructure]]></category>
            <category><![CDATA[logistics]]></category>
            <category><![CDATA[crowdgathergo]]></category>
            <dc:creator><![CDATA[Atif Hussain]]></dc:creator>
            <pubDate>Sat, 29 Aug 2026 04:46:15 GMT</pubDate>
            <atom:updated>2026-08-29T04:46:15.045Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZW2wqKw5RVXMSuUElr3Kag.png" /></figure><h3>How a stopping train can stop costing the network.</h3><p>Every plan for a saturated corridor ends in the same place. Lay another line.</p><p>Sometimes that is right. Often it is not the first thing that is short. Before you buy land, it is worth asking what is actually consuming the capacity you already own.</p><p>Two things are. Speed heterogeneity, and the stopping train.</p><h3>Why mixed speeds cost so much</h3><p>Put a superfast behind a passenger train on the same running line and you have not just slowed the superfast. You have spent the whole section at the slower speed, for both.</p><p>Headway is the distance you must keep behind the train ahead. When the trains ahead and behind run at different speeds, that distance keeps changing, so the chart has to carry padding for the worst case. Padding is capacity you have paid for and cannot use.</p><p>The moment every train on a section runs at the same speed, the geometry becomes stable, and close headway becomes possible.</p><h3>Defining the slot</h3><p>So we define the slot on the track, not on the train.</p><p>On a double line, take one directional track. Define slots that run at that section’s non-stop track speed. Space them five minutes apart. Every train type uses them: passenger, express, superfast, goods.</p><p>A slot is not an aspiration. It is a specific path, in space and time, that a train head occupies at each milestone.</p><h3>Then handle the stop honestly</h3><p>Which leaves the real problem. Trains stop. That is the point of most of them.</p><p>Today, a train that halts on the through line holds the block section while it decelerates, halts and accelerates. The train behind stands at a signal, waiting for track that is not moving anybody. Roughly six minutes of running time is lost to the braking and the climb back to line speed alone, before the halt itself is counted. About 5.6 minutes at 60 kmph, about 7.6 at 120.</p><p>Our answer is to stop treating the halt as an interruption of the run, and start treating it as an exit from the slot.</p><p>A train that plans to stop slows on approach and steps on to the loop line before it comes to rest. In doing so it gives up its slot. It halts clear of the running line. Trains behind take precedence and pass while it is standing — which is time it was going to spend stationary regardless.</p><p>It then stays only as long as it needs to be ready. Not a round two minutes. The actual number of seconds such that, once it picks up speed, it slots into the next free path cleanly. Then it goes.</p><p>It never waits at a signal for track availability.</p><h3>Overtaking stops being an event</h3><p>In a conventional working, an overtake is a decision. Somebody has to plan a crossing or precedence, hold one train, and accept the cost.</p><p>Under slot working, an overtake is simply what happens when a train stops. The stopping train is on the loop. The through line is clear. The following train runs through at line speed and nobody had to make a call.</p><p>That is the change worth arguing about. Not the headway number. The fact that stopping and overtaking became the same event instead of two competing ones.</p><h3>The arithmetic</h3><p>Five-minute slots, worked across the day with a three-hour maintenance block protected, gives about 250 slots a day on each directional track.</p><p>That number carries consequences beyond the count itself. Once paths are that predictable, spot ticketing and on-demand reservation become realistic. Freight moves without hunting for a window. And there is room to consider things the network cannot carry today, such as a passenger vehicle ferry, which would take road and short-haul air demand off the table for the same journey.</p><h3>Feasibility is arithmetic, not optimism</h3><p>None of this survives if the chart is aspirational.</p><p>So the chart is built at a granularity most working timetables do not use: to the second, and to the 100-metre milestone. Acceleration and braking are computed from kinematics, per train type, not assumed. The time at which a train head crosses each milestone is calculated, not padded.</p><p>If a slot cannot be occupied, the chart says so before the train is booked into it.</p><h3>The preconditions, stated openly</h3><p>Any officer reading this will probe, so here they are first.</p><p><strong>Elongated loops.</strong> A loop of about 1.5 km, so a train can enter and clear without dragging the section speed down. This is the main civil requirement, and it is small compared with a new line.</p><p><strong>Freight windowing.</strong> Goods run on the same slots, which means goods run to time. That is a working discipline change as much as a technical one.</p><p><strong>The safety-buffer question.</strong> For timetable changes that sit inside current safety-buffer thresholds, no new clearance regime is needed. The tighter case is real and worth naming: a stopping train’s gap shrinks to about 3.5 km as it enters the loop at 30 kmph, while the follower is still at full speed without two full blocks clear. Where that clashes with the current threshold, the honest offer is reduced slots on that section while RDSO evaluates the tighter regime. RDSO is the evaluator of the tighter case, not a gatekeeper on the ordinary one.</p><h3>Adherence: the other half</h3><p>A perfect chart that nobody can hold is a paper exercise.</p><p>WhatsComing shows a loco pilot where the train ahead and the train behind are on his track. It replaces guesswork with a picture. Lower mental load, fewer defensive slow-downs, better punctuality, and a safety gain that is independent of the timetable question.</p><h3>Where to start</h3><p>Not with the whole network. With one section, one direction, one worked Master Chart, and a microsimulation that anyone can interrogate.</p><p>Indian Railways has already run the ingredients separately. Surat-Vadodara did the prerequisites. The distinction matters and it is worth conceding plainly: that section did the groundwork, not this idea, and capacity stayed short, which is why the freight corridor followed. Apply the full idea and you get freight-corridor benefits without freight-corridor cost.</p><p>More trains. More punctual. The same tracks.</p><p><em>CrowdGatherGo — Innovations in eXtreme Mobility. UNCLOG — DISPERSE — MOVE. </em><a href="https://crowdgathergo.com/"><em>crowdgathergo.com</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=411ca9837fc3" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>