India has so many speed breakers because they are the cheapest, fastest way for a local body, a police station or a residents' group to slow traffic on mixed roads where pedestrians, carts and trucks share one carriageway. A correct Indian speed hump is defined in IRC:99-2018: for a 25 km/h crossing it is a circular arc with a 15 m radius, a 3.5 m chord and a 10 cm rise, painted and signed in advance. Most humps on Indian roads are shorter, steeper and unmarked, which is why ministry data linked speed breakers to 3,409 deaths in 2015 alone.
- A correct 25 km/h hump is 3.5 m long and 10 cm high, a 15 m radius arc (IRC:99-2018).
- It needs paint and a sign 40 m ahead; an unmarked hump is non-compliant.
- Lift-off speed is √(gR): about 44 km/h on the IRC hump, about 31 km/h on a typical short one.
- Speed breakers were linked to 3,409 deaths in 2015 (MoRTH).
- No humps on national highways; IRC prescribes 20 to 30 mm rumble strips instead.
The design problem was solved decades ago. The Indian Roads Congress (IRC) published its first speed-breaker code, IRC:99-1988, with a 17 m radius hump, 3.7 m wide and 0.10 m high, for an advisory speed of 25 km/h. The 2018 revision widened it into a full traffic-calming manual.
What failed was enforcement: humps are built by municipal engineers, contractors, panchayats and sometimes residents with a load of hot mix, and almost nobody checks them against the code.
This post explains what the code asks for, the physics of why the numbers matter, what the data says about harm, and how a correct hump would be built.
Why does India build so many speed breakers?
India builds speed breakers because its roads carry wildly different vehicles at wildly different speeds, and a hump is the only device that enforces itself without a policeman. IRC:99-2018 says so directly: Indian national and state highways pass through villages and towns, where tractors and animal carts share the carriageway with fast motorised traffic, and non-motorised users cannot be kept off highways.
The code's own speed numbers explain the appeal. It cites the Elvik power model (Elvik, Christensen and Amundsen, 2004): a 10% cut in mean speed reduces deaths by about 38%, while a 10% rise increases them by about 54%. It also notes that a 1 km/h fall in average speed typically cuts fatal crash frequency by about 3%. A hump delivers that reduction for the price of a few tonnes of bitumen, with no electricity, camera or staff.
Three practical forces multiply them:
- Local demand. After a crash outside a school or temple, residents ask for a hump, and a hump is visible proof that someone acted. Signs and painted bars feel like nothing has been done.
- Fragmented ownership. City roads, PWD roads, PMGSY rural roads and highways have different owners. The IRC code is a guideline, and each agency (and each contractor) builds to its own idea of a hump.
- Cheap repetition. When one hump does not slow drivers for long, the instinct is to add more. A field study in Hyderabad found humps only 20 to 30 m apart on one road, against an IRC spacing of 100 to 120 m.
The honking, cutting-in and lane indiscipline that make humps feel necessary are their own story; see why Indians honk so much.
At about 2 am on a Sunday in July 2026, a 28-year-old interior decorator was riding his motorcycle along the Outer Ring Road in north Bengaluru, near Chowdeshwari Nagar. The stretch was poorly lit, and the speed breaker on it was unlit. He hit it at speed and lost control; police said the bike was flung nearly 20 feet into the air before it landed back on the road. Passers-by rushed him to a nearby hospital with critical head injuries, but he did not survive. The Rajajinagar Traffic Police registered a case. The report does not say whether the hump was painted or signed, only that it was unlit on a dark road. Under the IRC code, markings and a sign 40 m ahead are part of the hump itself.
News Karnataka: Interior decorator killed after hitting unlit speed breaker on ORR (July 2026)
What does IRC:99-2018 say a speed breaker should look like?
IRC:99-2018 (Guidelines for Traffic Calming Measures in Urban and Rural Areas) specifies two hump shapes:
- Circular hump: defined by a radius and chord length with an assumed 10 cm rise.
- Trapezoidal (flat-top) hump: 50 to 100 mm high with ramps on both sides.
The radius or ramp length is chosen to match the speed you want drivers to cross at.
The code warns that circular humps lower than 10 cm let vehicles cross faster than the design speed, and humps higher than 10 cm may damage vehicles. It also says humps are not recommended on bus routes because of passenger discomfort; a longer speed table that fits a whole bus wheelbase is the alternative there.
| Design crossing speed | Circular hump radius | Chord length (10 cm rise) | Trapezoidal ramp length | Ramp gradient | Bus speed while crossing |
|---|---|---|---|---|---|
| 20 km/h | 11 m | 3.0 m | 0.7 m | 14.0% | 5 km/h |
| 25 km/h | 15 m | 3.5 m | 0.8 m | 12.5% | 10 km/h (circular), 5 km/h (trapezoidal) |
| 30 km/h | 20 m | 4.0 m | 1.0 m | 10.0% | 15 / 10 km/h |
| 35 km/h | 31 m | 5.0 m | 1.3 m | 7.5% | 20 / 15 km/h |
| 40 km/h | 53 m | 6.5 m | 1.7 m | 6.0% | 25 / 20 km/h |
| 45 km/h | 80 m | 8.0 m | 2.0 m | 5.0% | 30 / 25 km/h |
| 50 km/h | 113 m | 9.5 m | 2.5 m | 4.0% | 35 / 30 km/h |
Source: IRC:99-2018, Tables 3.1 and 3.2. If buses must pass at 40 km/h, the code asks for a 180 m radius and a 12 m chord.
Notice what that table implies: a correct hump is long. Even the slowest 20 km/h design is 3 m from edge to edge.
The familiar Indian hump, a sharp bump you can step over in one stride, is not a slower version of the IRC design. It is a different object.
What else does the code require: paint, signs and spacing?
A legal hump is invisible without markings, so the code treats paint and signs as part of the device. The 1988 code required:
- alternate black and white bands on the hump, preferably in luminous paint or with embedded cat's eyes;
- a "HUMP OR ROUGH ROAD" warning sign with a "SPEED BREAKER" plate 40 m before the first hump;
- repainting, which it called "the most important maintenance activity".
It also fixed spacing (100 to 120 m centre to centre when humps are repeated), asked that humps run across the full carriageway and shoulders so drivers cannot swerve around them, and banned humps on bridges. The 2018 code adds that road marking "shall be the essential part" of traffic calming, following IRC:35 for pavement markings, and that precast nail-down rubber bumps belong only on local and collector streets, because once broken their nails stay on the road and cut tyres.
Why are speed breakers banned on national highways?
Speed breakers are not allowed on national highways because a hump on a road designed for 80 to 100 km/h turns into a launch ramp for anyone who misses it. The road ministry told Parliament in 2017 that it had asked states and Union Territories not to build speed breakers on national highways, and that rumble strips may be used where speed control is unavoidable.
IRC:99-2018 matches that policy. Its rumble strips are raised bars 20 to 30 mm high, 200 to 300 mm wide and about one metre apart, roughly six per location, laid across the full carriageway including the shoulder. They rattle the vehicle and the driver's attention without throwing anyone off. Thermoplastic transverse bars, 5 mm or 15 mm high, do the same job with less vibration, and the number of sets increases with approach speed (three sets at 50, 80 and 120 m for 66 to 80 km/h).
The code also admits a limit: speed breakers on highways through villages have a zone of influence of only 70 to 80 m. Drivers brake for the hump and accelerate straight after it. For a whole village stretch, IRC prefers a graded speed zone: limits that step down from 60 km/h at 260 m from the boundary to 25 km/h inside it, with signs repeated 500 to 750 m out.
Courts keep returning to this. In July 2025, the Madhya Pradesh High Court declined to order removal of raised bars on NH-3 near Indore after finding they were rumble strips, not speed breakers. In August 2025, the Andhra Pradesh High Court directed the state to bring speed breakers on non-national-highway roads in line with IRC specifications.

What is the physics of a speed hump?
A circular hump works because a vehicle crossing it follows a curved path, and a curved path demands a vertical acceleration of \( a = v^2/R \). At the design speed that acceleration is a mild lift; above it, the lift grows with the square of speed until the wheels can leave the road. Here is the working, as I would check any hump.
- Find the radius from what you can measure. On the road you can measure the chord \(c\) (edge to edge) and rise \(h\). For a circular arc, \[ R = \frac{c^2/4 + h^2}{2h} \]For the IRC 25 km/h hump, \(c = 3.5\) m and \(h = 0.10\) m give \( R = (3.0625 + 0.01)/0.2 = 15.4 \) m, which matches the code's 15 m. The 1988 hump (17 m radius, 0.10 m rise) gives \( c = 2\sqrt{2Rh - h^2} = 3.68 \) m, its stated 3.7 m width.
- Vertical acceleration at the crest. At 25 km/h, the IRC design speed for that hump (\(v = 6.94\) m/s): \( a = 6.94^2/15.4 = 3.1 \) m/s², about 0.32 g. You feel it; nothing breaks.
- Lift-off speed. Contact is lost when \( v^2/R \ge g \), so \[ v_{\text{lift}} = \sqrt{gR} \]For the IRC hump, \( \sqrt{9.81 \times 15.4} = 12.3 \) m/s, or 44 km/h. Hit it at 50 km/h and the crest demands 1.28 g, more than gravity can supply.
- Now a typical short hump. One hump surveyed on Hyderabad's ECIL to NFC road measured 1.93 m wide and 0.06 m high (Korra Ravi Kiran et al., 2019). Its radius is \( (0.931 + 0.0036)/0.12 = 7.8 \) m and its lift-off speed is \( \sqrt{9.81 \times 7.8} = 8.7 \) m/s, about 31 km/h. A rider doing a normal 40 km/h city speed who sees it late reaches 1.6 g at the crest.
Calculated in this post from the IRC:99-2018 Table 3.1 hump (3.5 m chord, 0.10 m rise) and g = 9.81 m/s².
This is a point-mass model. A real motorcycle has tyres and suspension that absorb part of the rise, and two wheels that cross one after the other, so treat these as estimates of severity, not exact thresholds. The direction is robust, though: halve the radius and you halve the speed at which the hump becomes dangerous.
How harsh is this speed breaker?
How many people do badly built speed breakers kill?
The road ministry's own data linked speed breakers to 11,008 crashes and 2,633 deaths in 2014, 11,084 crashes and 3,409 deaths in 2015, and 9,583 crashes and 3,396 deaths in 2016. That is roughly nine deaths a day, a figure the ministry put before Parliament in 2017.
Those numbers come from police records under the "road features" heading of the annual Road Accidents in India report, so they count crashes where a hump was recorded as the road condition, not proof that the hump alone caused the death.
Later editions of the report put more weight on potholes and overspeeding, and I could not find a published speed-breaker count for recent years, so the chart stops where the verified data stops. Potholes, the other "road feature" killer, were linked to 2,324 deaths in 2016 and 2,161 in 2023.
See the data as a table
| Speed breakers (deaths) | Potholes (deaths) | |
|---|---|---|
| 2014 | 2633 | |
| 2015 | 3409 | |
| 2016 | 3396 | 2324 |
| 2023 | 2161 |
Source: MoRTH data reported to Parliament (2014-15), Road Accidents in India 2016, and Road Accidents in India 2023 via CAG road-safety analysis · Gaps mean no verified figure was found, not zero. Recent MoRTH reports do not publish a separate speed-breaker series that I could confirm.
For scale, India recorded 4,80,583 road crashes and about 1.72 lakh deaths in 2023, the highest toll on record. Speed breakers are a small share of that, but a uniquely avoidable one: they are objects that public money put on the road. The state-by-state picture is in road deaths by state, and for riders specifically, two-wheeler road deaths and touring.
Single cases show the pattern. In July 2026 a 28-year-old motorcyclist died after hitting an unlit speed breaker on a poorly lit stretch of the Outer Ring Road in north Bengaluru around 2 am, the case told at the top of this post; police registered a case. The story is common enough that every city has its own version.
The Osmania University team measured humps on two Hyderabad roads (a 3 km campus road and a 2 km ECIL to NFC stretch) with a radar gun, recording speeds 100 m before each hump and at the crest for two-wheelers, cars, autos, LCVs and heavy vehicles. Humps ranged from 1.93 to 3.02 m wide and 6 to 11.5 cm high, and none matched the IRC standard. Speed reduction rose with height and fell with width; a regression on the two dimensions explained 78% of the variation. Two-wheelers slowed by 49 to 60%. On the unmarked road, drivers only noticed humps about 50 m out and braked hard, while on the signed road they began slowing 100 m out. Simulation showed the humps dropping level of service from A to E on one road and C to E on the other. The limit: two roads in one city, and the regression is on a small set of humps.
Read the paper (free)The follow-up study looked at what humps cost road users. A test car run on a stretch with 1.66 speed breakers per km used about 68% more fuel than on a free-flowing stretch, and simulation in VISSIM showed the level of service falling from A without humps to E with them. The authors also regressed hump geometry against speed. The limit: the fuel result comes from a single test car, so treat 68% as an indication of scale for stop-start humps, not a national average.
Read the paper (free)How would you build a speed breaker properly?
A proper speed breaker starts with the question of whether a hump is the right tool at all; then it is built to the IRC profile, marked and signed. Here is the sequence the two IRC codes lay out, condensed.
- Check the warrant. IRC:99-1988 justified humps mainly on minor roads meeting highways (placed about 10 m from the major road's edge), mid-block urban sections, and residential, school, campus and hospital streets. Not on national or state highways, not on bus routes, not on bridges.
- Pick the design speed. Usually 20 to 30 km/h near schools and markets; IRC:99-2018 suggests 25 km/h near schools.
- Read the profile off the table. For 25 km/h: a circular arc of 15 m radius, 3.5 m chord, 10 cm rise; or a flat-top hump with 0.8 m ramps at 12.5%.
- Lay it full width. Across the carriageway and shoulders, with drainage gaps planned so water does not pond against it.
- Build to a template. The 1988 method: mark and roughen the surface, apply tack coat, set forms to the profile, pour and compact hot premix bitumen, trim edges, cure before opening.
- Mark and sign it. Black and white (or retroreflective) bands on the hump, cat's eyes if possible, and a "HUMP" warning sign with a "SPEED BREAKER" plate 40 m before it.
- Space and maintain. 100 to 120 m apart if repeated. Repaint before every monsoon, because a hump that has lost its paint is an unmarked hump again.
Do speed breakers make roads safer at all?
Yes, where they belong: a properly profiled hump on a residential or school street does cut speeds, and lower speed is the single strongest lever on crash severity. The Hyderabad data shows the same humps cut 85th-percentile speeds by about half. What the evidence does not support is using humps as a substitute for road design on fast roads, where their effect fades within 70 to 80 m and their failure mode is a crash.
The better toolkit, in IRC's own list, includes lane narrowing, raised pedestrian crossings (speed tables), stone-set or textured surfaces, chicanes, roundabouts and graded speed zones. They cost more to design and they need someone to think about the whole street. The deeper reason India has so many humps is that a hump is the one tool that needs no plan. Traffic as a whole is the subject of why Bengaluru traffic is so slow, and the suspension side of hitting one is in motorcycle suspension sag setup.

What's new
- Andhra Pradesh High Court orders speed breakers on state roads brought in line with IRC specifications LiveLaw, Aug 2025
- Madhya Pradesh High Court treats NH-3 bars as rumble strips and declines to order removal Verdictum, Jul 2025
- Pothole crashes killed 2,161 people in 2023, up 16% on 2022, per MoRTH data CAG, Oct 2025
- Motorcyclist killed after hitting an unlit speed breaker on Bengaluru's Outer Ring Road News Karnataka, Jul 2026
Questions people ask
What is the standard height of a speed breaker in India?
The IRC circular hump assumes a 10 cm rise; trapezoidal humps are 5 to 10 cm. The code says rises above 10 cm may damage vehicles. The length matters as much: 3.5 m edge to edge for a 25 km/h hump.
Are speed breakers allowed on national highways?
No. The road ministry has asked states not to build them on national highways, and IRC:99-2018 recommends rumble strips (20 to 30 mm bars) where speed control is unavoidable.
Why are so many speed breakers not painted?
Paint wears off in months and repainting is a maintenance budget line that often lapses. Both IRC codes treat markings and a warning sign 40 m ahead as part of the device, so an unpainted hump is technically non-compliant.
What is the difference between a speed bump and a speed hump?
In IRC usage, a hump is a long bitumen or concrete profile (3 m or more), while a bump is a short precast unit nailed to the road. The 2018 code limits bumps to local and collector streets because they become a hazard at speed and when broken.
Can I complain about an illegal speed breaker?
Yes. Write to the road-owning agency with photos, measurements and the IRC:99-2018 reference. In 2025 the Andhra Pradesh High Court ordered the state to align humps on its roads with IRC norms after such a petition.
Sources
- Indian Roads Congress: IRC:99-2018, Guidelines for Traffic Calming Measures in Urban and Rural Areas (2018; hump tables, rumble strips, speed zones, Elvik table)
- Indian Roads Congress: IRC:99-1988, Tentative Guidelines on the Provision of Speed Breakers for Control of Vehicular Speeds on Minor Roads (1988; 17 m radius, 3.7 m, 0.10 m, signs at 40 m, spacing)
- Korra Ravi Kiran et al.: Analysis of Speed Profiles at Speed Hump under Various Dimensions (WCTR 2019) (2019)
- Korra et al.: Impact Analysis of Speed Humps on Indian Roads, Journal of EASTS (2019)
- Swarajya: Nine lives lost daily in accidents due to speed breakers, Parliament told (2017; 2014 and 2015 MoRTH figures)
- Governance Now: Road Accidents in India 2016 figures (2017; 2016 speed breaker and pothole figures)
- CAG: Black spots and safe systems (2025; 2023 pothole figures from MoRTH)
- The Week (PTI): 1.72 lakh people died in 4.80 lakh road accidents in 2023 (2025)
- LiveLaw: Andhra Pradesh High Court on speed breaker specifications (2025)
- Verdictum: Tanishq Patel v NHAI, Madhya Pradesh High Court (2025)
- News Karnataka: Motorcyclist killed after hitting unlit speed breaker on ORR (2026)
