Two roads can carry the same number of lanes, be built to the same thickness and be signposted for the same speed, and still be entirely different roads. The difference is whether a tractor can join at any field boundary.
On 21 September 2026 the National Highways Authority of India released guidelines for the planning and design of access-controlled high-speed National Highways, other than expressways. That qualifying phrase is not bureaucratic hedging. It identifies a category of road that India has been building for years without a standards manual of its own, and the guidelines exist to close exactly that gap.
The category that had no rulebook
India's road standards are written by the Indian Roads Congress, and they are organised by road type.
Expressways have their own complete manual — IRC:SP:84, the Manual of Specifications and Standards for Expressways, alongside IRC:SP:99. An expressway in Indian practice is a divided, fully access-controlled carriageway with grade-separated interchanges, a design speed of 120 km/h, and no direct access from adjoining land. Ordinary National Highways have their own long-established standards, built around a design speed nearer 100 km/h and intersections at grade.
Between those two sits a road that is neither. It is a four- or six-lane National Highway built with access control and intended for high speed, but not designated or specified as an expressway. NHAI has been procuring these corridors, and the consultants preparing Detailed Project Reports for them have been improvising — borrowing from the expressway manual where it seemed to fit and from ordinary highway standards where it did not. The result, in NHAI's own words, was observable variation between projects.
The new guidelines apply to all upcoming four- and six-lane greenfield and brownfield high-speed access-controlled National Highways. Their purpose is uniformity: the same corridor type should be designed the same way whichever consultant prepares the DPR and whichever state it runs through.
What access control actually does
The phrase sounds administrative. It is a piece of engineering with a measurable safety mechanism behind it, and the mechanism is the conflict point.
A conflict point is any location where vehicle paths cross, merge or diverge. Count them at a conventional four-leg intersection of two-lane roads and the total is 32: eight merging, eight diverging, and — the dangerous ones — sixteen crossing conflicts, where one vehicle's path passes directly through another's at an angle.
Crossing conflicts are what produce right-angle and head-on collisions, the crash types that kill. At 40 km/h a right-angle impact is survivable in a modern car. At 100 km/h it is generally not, because the kinetic energy that the structure must absorb rises with the square of the speed: doubling from 50 to 100 km/h quadruples the energy. This is the engineering reason speed and access control cannot be separated. A road designed for high speed with uncontrolled access is not a fast road — it is a road where the consequences of an ordinary mistake become fatal.
Access control addresses this at the source. IRC:62, the standing guideline on control of access, distinguishes full control from partial control. Under full control, the highway is a closed system: vehicles enter and leave only at points the designer chose, and the owners of land along the route have no right of direct access onto the carriageway. Their frontage right is extinguished and replaced by a service road or an indirect connection. Under partial control, some direct access survives at selected locations.
Grade separation then removes the crossing conflicts entirely. If the crossing road passes over or under rather than through, the sixteen most dangerous conflict points cease to exist. What remains at the interchange are merging and diverging movements — conflicts that occur at low relative speed and at shallow angles, which is precisely why they are survivable.
Widths matched to traffic
The guidelines standardise lane configuration against a detailed traffic assessment, with the right of way — the width of land acquired for the corridor — following from it.
| Projected traffic | Configuration | Width |
|---|---|---|
| Up to 15,000 vehicles | 4-lane highway, 4-lane structures | 60 metres |
| 15,000–25,000 vehicles | 4-lane highway, 6-lane structures | 70 metres |
| 25,000–40,000 vehicles | 6-lane highway, 6-lane structures | 70 metres |
Two things in that table are worth noticing.
The first is that structures are built wider than the carriageway they carry. In the middle band the highway is four lanes but the bridges and viaducts are six. Widening a road is comparatively straightforward — the land is already acquired and the pavement is extended. Widening a bridge is not; it is close to rebuilding it. Designing structures one step ahead of current traffic is how a corridor is made expandable without demolition later.
The second is that right of way is acquired for the eventual road, not the initial one. Land acquisition is the slowest, most contested and most expensive part of any Indian highway project. Acquiring 70 metres once and building six lanes into it over time is cheaper and faster than returning to the same landowners a decade later.
Traffic in such assessments is counted in passenger car units. A PCU expresses a vehicle's effect on traffic flow relative to a standard car, which is 1.0 PCU. A truck or a bus occupies more road space and accelerates more slowly, so it counts as several PCUs; a two-wheeler counts as a fraction. Converting a mixed Indian traffic stream into PCUs is what makes a capacity calculation meaningful, and it is the reason two roads carrying the same number of vehicles can need different numbers of lanes.
Where the interchanges go
The guidelines specify that interchanges with ramps be provided where the corridor crosses expressways, National Highways, State Highways and important arterial roads generating substantial traffic, and at roads connecting ports, airports, Multi-Modal Logistics Parks, industrial areas and major tourist destinations.
That list is the logistics network written into the road design. An access-controlled corridor delivers its benefit only where traffic can actually get on and off it; an interchange in the wrong place leaves a port or an industrial cluster stranded a few kilometres from a high-speed road it cannot use. Planning interchanges around freight generators rather than around settlements is a deliberate choice about what the corridor is for.
There is a rough capacity rule underlying the decision. An at-grade intersection begins to fail when demand exceeds roughly 10,000 PCU per hour, at which point an interchange becomes necessary rather than merely desirable. Interchanges are expensive — they consume land, require structures, and cost many times an at-grade junction — so their spacing is always a compromise between access and economy. Too few and the corridor serves nobody along its length; too many and it stops behaving like a high-speed road at all.
The wall, and the sign before the exit
Two smaller provisions carry more engineering than their size suggests.
The first is a 1-metre-high boundary wall along the outer edge of the corridor. A metre is not a crash barrier — vehicle restraint is a separate system, placed in the median and at hazards, and designed to redirect a vehicle rather than stop it dead. This wall does a different job: it physically enforces the access control. It keeps out encroachment, unauthorised entry, and stray cattle. On Indian highways an animal on the carriageway at night is a genuine and recurring cause of high-severity crashes, and at 100 km/h a driver's reaction plus braking distance is longer than the distance at which an unlit animal becomes visible. The wall removes the hazard rather than asking the driver to cope with it.
The second is the requirement for advance notification of exits and entries. This is human-factors engineering, and its principle is driver expectancy. A driver on a high-speed road needs to know what is coming far enough ahead to decide, check mirrors, change lanes and decelerate — a sequence that takes time, and at 100 km/h covers a great deal of ground. Traffic engineering calls the distance needed for this decision sight distance, and it is substantially longer than simple stopping sight distance. Late signage produces the most dangerous manoeuvre on any access-controlled road: the sudden multi-lane cut across traffic by a driver who has seen the exit too late. Signing the exit early is not courtesy. It is the countermeasure for a specific and well-documented crash type.
The number behind the guidelines
All of this sits under a stated target: 50,000 km of access-controlled high-speed corridors under Viksit Bharat 2047.
The rationale offered is logistics cost. Freight moving at a predictable speed over a corridor with no at-grade interruptions is cheaper per tonne-kilometre than freight on a road where every town imposes a stop — not only because the journey is faster, but because the journey time is reliable, and reliability is what lets a firm hold less inventory. Our explainer on multi-lane free flow tolling covers the complementary half of this: a corridor with no at-grade junctions is wasted if it still has vehicles stopping at toll plazas.
The same logic runs down the rest of the network. Rural connectivity under PMGSY works on access rather than speed, and the materials research described in our piece on steel slag road technology applies to both. A highway network is a hierarchy, and each level is designed to do something the others cannot.
Whether 50,000 km is delivered on time is a separate question from whether it is designed consistently. The guidelines address only the second. But a standards document is not a minor instrument — it is the point at which a policy target becomes a set of instructions that thousands of engineers will actually follow.
🔑 Revision block
- Issued: NHAI, 21 September 2026 — guidelines for planning and design of access-controlled high-speed National Highways other than expressways
- Applies to: all upcoming 4- and 6-lane greenfield and brownfield high-speed access-controlled NHs
- Target: 50,000 km of access-controlled high-speed corridors under Viksit Bharat 2047
- Gap filled: expressways have IRC:SP:84 and IRC:SP:99; this intermediate category had no separate standard, producing DPR variation
- Access control: IRC:62 distinguishes full control (no direct access from abutting land) from partial control
- Conflict points: conventional four-leg at-grade intersection has 32 — 16 crossing, 8 merging, 8 diverging
- Widths: up to 15,000 vehicles → 4-lane highway with 4-lane structures, 60 m; 15,000–25,000 → 4-lane with 6-lane structures, 70 m; 25,000–40,000 → 6-lane with 6-lane structures, 70 m
- Interchanges: at expressways, NHs, SHs, major arterials, and roads to ports, airports, Multi-Modal Logistics Parks, industrial areas, tourist destinations
- Boundary wall: 1 metre high along the outer edge — access control and stray cattle, not crash protection
- Signage: advance notification of exits and entries, addressing decision sight distance
- Design speed reference: expressway 120 km/h; ordinary 4-lane NH nearer 100 km/h
- PCU: passenger car unit; standard car = 1.0; at-grade intersections fail beyond roughly 10,000 PCU per hour
🎯 Practice MCQs
Q1. The NHAI guidelines issued on 21 September 2026 apply to: (a) Access-controlled high-speed National Highways other than expressways (b) Expressways only (c) Rural roads under PMGSY (d) State Highways
→ (a) — expressways already have IRC:SP:84.
Q2. How many conflict points does a conventional four-leg at-grade intersection of two-lane roads have? (a) 16 (b) 24 (c) 32 (d) 48
→ (c) — 16 crossing, 8 merging, 8 diverging.
Q3. Grade separation at an intersection primarily eliminates: (a) Merging conflicts (b) Diverging conflicts (c) All conflicts of every kind (d) Crossing conflicts
→ (d) — merging and diverging movements remain at the ramps, but they occur at shallow angles and low relative speed.
Q4. Under full control of access as defined in Indian road standards: (a) Vehicles may enter anywhere but must exit at designated points (b) Owners of abutting land have no right of direct access onto the carriageway (c) Only commercial vehicles are restricted (d) Access is controlled only at night
→ (b)
Q5. A corridor with projected traffic between 25,000 and 40,000 vehicles is to be built with: (a) 4-lane highway and 4-lane structures, 60 m width (b) 4-lane highway and 6-lane structures, 70 m width (c) 6-lane highway and 6-lane structures, 70 m width (d) 8-lane highway, 90 m width
→ (c)
Q6. Structures are specified wider than the carriageway in the middle traffic band because: (a) Widening a bridge later is far costlier and more disruptive than widening a road (b) Bridges must be heavier than roads (c) It reduces construction time (d) Regulations require symmetrical design
→ (a) — designing structures one step ahead makes future expansion possible without rebuilding.
Q7. One passenger car unit (PCU) corresponds to: (a) One person travelling (b) One standard passenger car (c) One tonne of freight (d) One kilometre travelled
→ (b) — heavier and slower vehicles count as more than 1.0; two-wheelers as less.
Q8. The 1-metre boundary wall prescribed along the corridor edge is primarily intended to: (a) Restrain vehicles that leave the carriageway (b) Reduce noise for nearby settlements (c) Carry drainage (d) Prevent encroachment, unauthorised access and entry of stray cattle
→ (d) — vehicle restraint is a separate system with different design requirements.
Q9. "Decision sight distance" differs from stopping sight distance in that it: (a) Accounts for the time needed to perceive, decide and execute a manoeuvre, not merely to brake (b) Is shorter, because drivers react faster on highways (c) Applies only in urban areas (d) Is measured at night only
→ (a) — which is why advance notification of exits is a safety provision rather than a convenience.
Q10. Consider the following statements: 1. Access-controlled corridors reduce logistics costs partly by making journey times predictable. 2. An interchange is required when an at-grade intersection can no longer handle the traffic demand. Which is/are correct? (a) 1 only (b) 2 only (c) Both 1 and 2 (d) Neither 1 nor 2
→ (c) — predictability lets firms hold less inventory, which is a cost saving distinct from speed.
📋 How this gets asked (PYQ pattern)
Infrastructure questions in CDS and OTA tend to be factual, but the road sector is one place where a conceptual question appears with some regularity, and candidates who have only memorised scheme names find it hard.
The classification question asks what distinguishes an expressway from a National Highway. The answer is access control and design speed, not lane count. A six-lane National Highway with at-grade junctions is not an expressway; a four-lane fully access-controlled road designed for 120 km/h is closer to one. Candidates who answer by counting lanes get this wrong.
The standards question asks which body writes India's road specifications. The Indian Roads Congress. It appears often enough to be worth holding, along with the fact that expressways have a dedicated manual.
The authority question separates NHAI, the Ministry of Road Transport and Highways, NHIDCL and the Border Roads Organisation. NHAI develops and manages National Highways; NHIDCL works in the North East and border areas; BRO builds and maintains strategic roads in frontier regions.
The concept question is the one worth preparing deliberately: conflict points, grade separation, right of way, passenger car units. These are defined terms in traffic engineering, and they carry across into any question about road capacity or safety.
For the descriptive paper, the useful frame is that a road network is a hierarchy with different design objectives at each level — access at the rural end, mobility at the corridor end — and that the intermediate categories are where most of the design difficulty lies. An answer that explains why a country needs more than one class of highway is stronger than one that lists schemes and targets.
Preparing for CDS or OTA? Infrastructure answers improve sharply once you can explain a mechanism rather than quote a target. Build the base with our CDS/OTA general studies notes, follow the daily CDS/OTA current affairs, and prepare with our faculty in the upcoming Cavalier courses in Delhi.
✍️ Written by Aditya Tiwari — Economy & polity faculty at The Cavalier. Reviewed by the Cavalier Faculty Desk.