Based on MIT AVT Consortium naturalistic driving research, MIT CSAIL pedestrian behaviour modelling,
EU Road Safety Observatory VRU data, WHO pedestrian safety, and RSA Ireland VRU statistics
Vulnerable Road Users (VRUs) are road users who lack the physical protection of an enclosed metal vehicle. They are exposed to the full energy of any collision without the crumple zones, airbags, and structural protection that car occupants rely on. VRUs include pedestrians, cyclists, motorcyclists, e-scooter users, horse riders, and people with mobility aids. They represent a disproportionate share of road fatalities in every country in the world.
Completely unprotected. Any collision with a vehicle at speed is potentially fatal. Most vulnerable: children, older adults, people with disabilities. Most at risk: urban crossings, night time, alcohol-impaired pedestrians.
Ireland: Pedestrians account for ~21% of road fatalities but less than 5% of vehicle-km travelled.
Helmet provides some head protection; body is fully exposed. Vulnerable to: dooring (car doors opening), being squeezed at junctions, being hit from behind, and falls from road surface defects.
Ireland: Cycling has grown significantly but cycling infrastructure has not kept pace — fatality rate per km is 10× that of car occupants.
Maximum personal protective equipment (PPE) reduces injury severity but cannot protect against the physics of high-speed collision. Motorcycle crashes are 29× more likely to be fatal than car crashes per km travelled.
Ireland: Motorcyclists are 15–20% of road fatalities while representing ~2% of vehicle-km.
Understanding what physically happens to a pedestrian in a collision helps explain why speed is so critical. The injury pattern in a pedestrian collision is determined by the vehicle's speed, height, and front-end geometry — and the pedestrian's height, age, and body mass. MIT AVT research on vehicle-pedestrian interaction has directly influenced EU pedestrian protection regulations and front-end design standards.
The vehicle bumper strikes the pedestrian's legs (or pelvis/abdomen in children). The pedestrian's lower body is stopped; upper body continues forward due to inertia. This creates a "wrap" effect — the pedestrian bends forward onto the bonnet.
The pedestrian's head strikes the bonnet, windscreen, or A-pillar. This is often the most severe and fatal injury phase. Head impact speed is approximately 1.5× vehicle speed because the pedestrian's body adds rotational velocity. A car at 50 km/h produces a head impact equivalent to 75 km/h.
The pedestrian is thrown forward (or sideways) and impacts the road surface. Secondary injuries: additional head trauma, spinal injury, friction burns. If vehicle braked hard, pedestrian may be run over.
At very high speeds, the pedestrian may be dragged under the vehicle. Survivability is very low.
The dramatic fall in pedestrian survival between 30 and 50 km/h is explained by kinetic energy (KE = ½mv²). Speed increase from 30→50 km/h increases KE by 278%. The human skull can absorb approximately 80–100 J of impact energy. A 50 km/h head impact typically delivers 200+ J.
MIT AVT researchers modelled pedestrian head impact trajectories for different vehicle geometries. Their data showed that SUVs and high-bonnet vehicles produce higher head impact severity than traditional sedan profiles — because the taller front end strikes the pedestrian higher on the body, transferring more energy to the torso before head contact. This research influenced EU pedestrian safety test protocols for high-bonnet vehicles, now including separate SUV pedestrian impact tests in Euro NCAP.
A child's centre of mass is higher relative to their height. A bumper that hits an adult at knee level hits a 5-year-old at the pelvis or chest. Primary impact for children is more severe — causing internal organ injury before secondary head impact. Children's pedestrian injuries have a different anatomical profile to adults and require different protective measures.
MIT CSAIL has developed some of the world's most advanced pedestrian behaviour prediction models — originally for autonomous vehicle safety systems, but with direct implications for human driver awareness. Their research reveals that pedestrian behaviour follows predictable patterns — and that human drivers consistently fail to anticipate these patterns correctly.
Using computer vision analysis of thousands of pedestrian-vehicle interactions, MIT CSAIL identified the key variables pedestrians use to decide whether to cross:
Critical finding: Pedestrians consistently underestimate vehicle speed by 10–20%. A car doing 60 km/h looks like 48–50 km/h to a crossing pedestrian.
MIT CSAIL research on pedestrian-driver interaction found that both parties are trying to read each other's intentions — but the communication is ambiguous and often misread. Pedestrians watch for: eye contact, slowing vehicle, vehicle body lean (indicating braking). Drivers watch for: pedestrian looking at them, pedestrian stepping back or forward, hand gestures.
When both parties are simultaneously uncertain, both may proceed — with catastrophic consequences. The car always wins the physics. This research directly motivated development of AV communication displays ("It is safe to cross" messages to pedestrians).
When a vehicle stops to let a pedestrian cross, the pedestrian may step out into the adjacent lane without checking — hidden from the next vehicle by the stopped one. Multiple threat is the most common scenario in complex pedestrian fatalities. Always check all lanes before proceeding, even when another vehicle has stopped.
Cyclist fatalities and serious injuries are disproportionate to cyclist numbers. Understanding the specific mechanics of cyclist crashes — where they happen, how they happen, and what makes them survivable or not — reveals precisely what drivers need to do differently. Many of the worst cyclist crashes are entirely predictable and preventable.
Ireland's Road Traffic Act requires a minimum 1.5m lateral clearance when overtaking a cyclist. This distance is based on biomechanical research showing:
At 60+ km/h, the safe passing distance increases further. On narrow roads: wait, do not squeeze past.
Research shows most drivers open car doors without checking for cyclists. The Dutch Reach method — opening the door with the far hand, forcing the body to turn and see approaching traffic — reduces dooring by creating a natural head-check. Standard in Dutch driver training. Now being introduced in Irish EDT curriculum. At 25 km/h, a cyclist cannot stop in time for a suddenly opened door at less than 3m distance.
Heavy goods vehicles have large blind spots to the left side and front. A cyclist alongside an HGV at a junction is invisible to the driver. Never pass an HGV on the left — move ahead of it or wait behind. This rule has been taught in cyclist training for years but not driver training. EU Direct Vision Standard (2019) mandates improved cab visibility in new HGVs to address this.
"Sorry mate, I didn't see you" is the single most common explanation given by drivers who collide with motorcyclists — and it is genuinely true. The driver looked, but did not perceive the motorcycle. SMIDSY is not an excuse — it is a predictable consequence of how human visual perception works, combined with the unique profile of a motorcycle. Understanding the mechanism is the first step to preventing it.
The human visual system detects objects through contrast, motion, and size. A car approaching a junction is wide, high-contrast, and optically large. A motorcycle is narrow (0.7m vs. 2.1m for a car), appears smaller, and has a different speed-size relationship. At the same distance, a motorcycle looks further away than a car.
A motorcycle approaching head-on shows minimal angular motion — it appears to grow in size (looming) rather than moving laterally. The human visual system evolved to detect lateral movement (escaping prey, approaching threats from the side). Head-on objects moving in a straight line are neurologically less salient — they are harder to detect as "approaching fast."
The driver's visual system is scanning for "car-sized" objects. The motorcycle does not match the stored neural template for "approaching vehicle." The brain literally fails to flag it as a threat, even when the eyes point at it. This is inattention blindness driven by expectation mismatch.
Because motorcycles appear smaller, drivers underestimate their speed. A motorcycle at 80 km/h is judged to be travelling at 55–65 km/h. The driver judges the gap to be safe — but it is not.
MIT CSAIL's computer vision research quantified visual salience (how attention-grabbing an object is) for different road users under different conditions. Motorcycles scored significantly lower salience than cars in urban junction contexts — particularly when approaching in bright sunlight or against complex backgrounds. This research informs AV perception system design and is being used to develop enhanced motorcycle detection algorithms in ADAS systems.
Conspicuity aids (bright clothing, reflective material, daytime running lights) help but do not solve the fundamental perception problem. The best motorcyclist protection is driver education about SMIDSY — and road design that reduces junction conflict.
Motorcycle crashes are 29× more fatal per km than car crashes. Understanding why — through the physics of energy transfer and the biomechanics of rider injury — reveals what protective equipment can and cannot do, and why road design and driver awareness remain the most important safety factors.
A motorcycle and rider at 80 km/h carry kinetic energy of approximately 80 kJ. In a collision, this energy must be absorbed — by deformation of the motorcycle, by the road surface (friction), or by the rider's body.
Unlike a car, a motorcycle has no crumple zone protecting the rider. The rider is the crumple zone. At speeds above 60 km/h, no amount of PPE can fully protect against this energy transfer to the human body.
| PPE | Injury Reduced | Effectiveness |
|---|---|---|
| Full-face helmet | TBI, facial injury | 42% reduction in fatality |
| Back protector (CE-A) | Spinal injury | Significant in low-medium speed falls |
| Airbag system | Thorax, neck, shoulder | Up to 60% impact force reduction |
| Abrasion-resistant jacket | Road rash, friction burns | Prevents major skin/tissue loss |
| Motorcycle boots | Ankle, foot fracture | Significant reduction in ankle injury |
MIT AVT Consortium researchers evaluated Automatic Emergency Braking technology adapted for motorcycles. Unlike car AEB, motorcycle AEB cannot simply apply maximum braking — it must modulate braking to prevent wheel lock and highside falls. MIT AVT research contributed to the development of cornering ABS and IMU-based braking systems that reduce motorcycle crash severity while maintaining stability.
Children are not small adults. They interact with the road environment in ways that are fundamentally different from adults, and drivers need to understand these differences to anticipate child pedestrian behaviour. MIT CSAIL and EU child road safety research shows that children's cognitive limitations create predictable hazard patterns that drivers can learn to recognise.
Peak risk time: Monday–Friday, 3–6pm, in residential streets within 300m of a school.
While Safe System thresholds set 30 km/h as the general unprotected-pedestrian survival threshold, research specifically on child pedestrian crashes recommends 20 km/h in areas where children are present. A child's smaller body mass and head size means they sustain more severe injuries at the same impact speed as adults. 20 km/h speed limits outside schools reduce child pedestrian fatalities by up to 70% (Swedish and Dutch evidence).
The most common child pedestrian crash scenario: child runs between parked cars. The car driver has no warning — the child appears in the road with zero time to stop. At 30 km/h, stopping distance (reaction + braking) is approximately 14m. If the child appears at 8m, the crash is unavoidable at 30 km/h. At 20 km/h, the crash is survivable. This is why residential speed limits matter.
Older adults are over-represented in pedestrian fatalities relative to their share of pedestrian trips. MIT AgeLab's research on ageing and mobility provides a detailed picture of why older pedestrians face elevated risk — and what road design changes and driver behaviours would protect them.
MIT AgeLab researchers studied older pedestrian behaviour at signalised crossings across Boston. They found that 34% of older adults over 75 were unable to complete a standard crossing phase in the time allocated. Many started crossing in green and were still crossing at red — with vehicles legally beginning to move. Recommendation: pedestrian signal phases should use a minimum of 1.0 m/s walking speed calculation for all crossings.
At junctions, give way to older pedestrians even if the signal has changed or they are still crossing — they cannot accelerate to clear the road as younger people can. Patience is legally required and ethically essential. Driving around them is a serious road offence and a moral failure.
Night-time driving represents roughly 25% of all vehicle-km in Ireland but accounts for approximately 40% of fatal crashes. For VRUs, the night-time risk multiplier is even more severe. A pedestrian in dark clothing at night may not be detectable by a driver until they are within stopping distance at even moderate urban speeds — making night-time pedestrian crashes among the most difficult to prevent through driver reaction alone.
A driver's ability to detect a pedestrian depends on:
Dark clothing: Detection at 25–30m with dipped headlights
Light clothing: Detection at 55–65m with dipped headlights
Reflective material: Detection at 120–140m
High-viz vest: Detection at 150–180m
Stopping distance at 50 km/h (wet): 36m
Stopping distance at 60 km/h (wet): 54m
Result: A driver at 60 km/h cannot stop for a pedestrian in dark clothing even if they react instantaneously at maximum detection range.
Driving faster than you can stop within your headlight range is "overdriving your headlights." At 100 km/h on dipped beams (illuminating ~60m), stopping distance is ~70m. You are always driving faster than you can see.
Main beam extends to 120–150m — but main beam must be dipped for oncoming vehicles, which is precisely when you need it most (approaching a vehicle = approaching any pedestrian near the road).
MIT AVT Consortium research evaluated automotive night-vision systems and pedestrian AEB performance at night. Findings: standard camera-based AEB systems operate at significantly reduced effectiveness in darkness compared to daylight (50–60% vs 80–90% of daylight performance). Radar-based AEB maintains effectiveness regardless of visibility. The research supported EU regulatory pressure for mandatory radar integration in AEB systems (not just camera).
38% of pedestrians killed in Ireland test positive for alcohol. Alcohol affects pedestrian visibility behaviour — impaired pedestrians step out without checking, wear dark clothing, walk in the road, and have reduced roadside awareness. The combination of impaired pedestrian + driver at speed after dark is the most common fatal pedestrian crash scenario in Ireland.
Most VRU deaths are predictable outcomes of infrastructure that was designed for vehicle movement without considering pedestrian and cyclist safety. Safe System infrastructure design treats VRU protection as a primary requirement — not an afterthought. Countries that have invested in VRU infrastructure have seen dramatic reductions in VRU fatalities without reducing mobility.
The Netherlands has the lowest cyclist fatality rate in the world despite very high cycling volumes. Key differences from Ireland: 35,000km of segregated cycle infrastructure (vs. ~2,000km in Ireland), filtered permeability (motor traffic cannot take short-cuts through residential areas), junction design that gives cyclists priority over motor traffic at low-speed junctions. Dutch cyclists are 5× safer per km than Irish cyclists.
ADAS technology specifically designed for VRU detection and protection represents one of the most promising developments in road safety. MIT AVT Consortium is at the forefront of research evaluating and improving these systems. The technology is advancing rapidly — but real-world effectiveness still falls short of laboratory performance in many conditions.
Automatic Emergency Braking systems with pedestrian detection use camera and/or radar to identify pedestrians in the vehicle's path and automatically brake if the driver has not responded.
Effectiveness: Euro NCAP testing shows P-AEB reduces pedestrian crashes by 38% in daylight conditions. Real-world effectiveness is lower — approximately 25% reduction.
Limitations: Night performance 50–60% of daylight. Crossing pedestrians are harder to detect than walking in-line. Cyclists are detected less reliably than pedestrians.
Extended AEB capable of detecting cyclists. Required in Euro NCAP from 2024. More challenging than P-AEB because cyclists have variable profiles, speeds, and trajectories.
MIT AVT research has contributed to improving C-AEB performance specifically for the left-hook scenario — detecting cyclists that would otherwise be in the driver's A-pillar blind spot during right turns.
Surround-view camera systems display a virtual bird's-eye view of the vehicle and its immediate surroundings. Originally developed for parking — increasingly used for low-speed VRU detection (delivery areas, school zones). Proximity warning alerts driver to objects within 1–2m.
Vehicle-to-Everything (V2X) communication allows vehicles to receive signals from connected pedestrians and cyclists via their smartphones. MIT's work in intelligent transport systems has demonstrated that a V2X-enabled pedestrian app can alert a vehicle when the pedestrian is about to cross — giving the driver 2–4 seconds more warning than camera detection alone. Pilot deployments in Singapore and Netherlands show 40% reduction in VRU near-miss events at equipped junctions.
All VRU detection systems have detection failure modes. P-AEB fails at night, in heavy rain, for unusual postures (crouching children, seated wheelchair users). AEB must be treated as a last resort — not a reason to disengage attention in areas where VRUs are present.
In law and in ethics, drivers bear a higher duty of care toward more vulnerable road users. This is not a bureaucratic rule — it is a recognition that operating a powerful machine near unprotected people creates an obligation to take extraordinary care. Understanding what this duty means in practice helps drivers make the right decisions in the marginal situations where casualties happen.
In Irish law, the driver is judged against the standard of a "reasonable and competent driver." Near VRUs, a reasonable driver does not merely obey the letter of the law — they reduce speed, increase attention, and allow for the possibility that the VRU will behave unpredictably.
A child who darts out: the reasonable driver near a school or residential area should already have reduced speed because children are reasonably foreseeable.
In traffic law, the party who had the last clear opportunity to avoid a collision may bear greater legal responsibility — even if the other party was initially at fault. A pedestrian who crosses illegally may be contributorily negligent — but a driver who had 3 seconds of detection time and did not brake may still bear primary liability.
This doctrine means: even when a VRU is doing something wrong, if you had the opportunity to avoid them and did not take it, you share legal and moral responsibility.
The legal minimum is the floor, not the ceiling. Drivers should ask not just "am I legally allowed to proceed?" but "is it safe to proceed?" Near VRUs these are often different questions. A driver who is legally in the right but hits a child who stepped out is legally defended but morally devastated.
The most experienced, safest drivers are not those who maximise their own rights — they are those who maximise safety for everyone sharing the road.
Speed is the single most important variable in VRU crash outcomes. Not because it is the most common cause of VRU crashes — but because it determines whether a crash is survivable. At the speeds common in urban environments, the difference of 10 km/h is the difference between a pedestrian who survives and one who does not.
| Impact Speed | Pedestrian Survival |
|---|---|
| 20 km/h | ~99% |
| 30 km/h | ~90% |
| 40 km/h | ~60% |
| 50 km/h | ~25% |
| 60 km/h | ~8% |
| 70 km/h | <3% |
Source: Rosen et al. (2011), based on 700+ pedestrian crash records
Stopping distance = reaction distance + braking distance.
| Speed | React(1.5s) | Brake(dry) | Total |
|---|---|---|---|
| 30 km/h | 12.5m | 4.5m | 17m |
| 50 km/h | 20.8m | 12.5m | 33m |
| 60 km/h | 25m | 18m | 43m |
An 8-year-old who runs between parked cars appears at approximately 5–8m. At any of these speeds, the crash cannot be avoided by braking alone. Speed reduction before the hazard zone is the only effective prevention.
50 km/h on a motorway is safe. 50 km/h in a school zone is potentially lethal. Speed only has meaning relative to the environment. A driver who travels at the "correct" speed for conditions — not just within the legal limit — is a safer driver. Look for VRU indicators: schools, playgrounds, residential streets, bus stops, pedestrian crossings — and reduce speed proactively.
The micromobility revolution — e-scooters, e-bikes, cargo bikes — has added new categories of VRU to the road environment. These users occupy an uncertain space in road design and law, travel at speeds between pedestrians and cyclists, and are less predictable to motor vehicle drivers trained before their widespread adoption. MIT transportation research has begun studying micromobility safety as a priority emerging area.
MIT Transportation Laboratory research on shared e-scooter deployments in US cities found that injury rates per trip are approximately 3× higher for e-scooters than for cycling. Key risk factors: shared footways (conflict with pedestrians), poor road surface (kerb strikes, pothole instability), and low helmet use. E-scooter users also show higher distracted-use rates than cyclists — partially attributed to the lower perceived effort of travel.
While VRU infrastructure discussions focus on urban areas, rural VRU crashes are disproportionately severe. Higher vehicle speeds, absence of footpaths, narrow lanes, no street lighting, and limited emergency response time combine to make rural VRU crashes more likely to be fatal than equivalent urban crashes.
RSA Ireland data shows that while most pedestrian crashes occur in urban areas, rural pedestrian crashes have a higher fatality rate per crash: 34% of rural pedestrian crashes are fatal vs. 8% of urban pedestrian crashes. The speed differential explains this entirely — urban pedestrian crashes typically involve vehicles at 30–50 km/h; rural crashes involve 80–100 km/h impacts.
Everything covered in this module translates into specific scanning behaviours, speed adjustments, and decision rules that every driver can apply on every journey. VRU awareness is not a special mode you switch on — it is an integrated part of attentive, expert driving at all times.
Advanced drivers approach every environment as if it might contain an unprotected road user who has not seen them. This is not pessimism — it is accurate probabilistic thinking. On an average urban journey, you will encounter dozens of VRUs. Most interactions are uneventful. But the VRU who makes an error while you are driving at 60 km/h is the one who may not survive. The precautionary speed reduction costs you seconds. The failure to reduce costs them their life.
In any situation where a vulnerable road user could be present — reduce your speed before the hazard zone, not when you see them. At 30 km/h, a pedestrian who steps out unexpectedly at 10m is survivable. At 50 km/h, they are not. You cannot see all VRUs in time to stop at urban speeds. The only protection is to have already slowed.
Ireland's VRU crash data reveals specific patterns that are both informative and actionable. Understanding where, when, and how VRU fatalities occur in the Irish context allows drivers, planners, and policymakers to target interventions precisely.
The RSA Safe System Strategy 2021–2030 includes specific VRU targets:
• 30 km/h speed limits in all urban areas by 2025
• 1,000km of new cycling infrastructure per year
• Protected crossings at all urban junctions above threshold traffic volume
• School 20 km/h zones at all schools by 2027
• Mandatory AEB with pedestrian/cyclist detection in new vehicles (EU mandate 2024)
In 2023, Ireland moved to implement 30 km/h as the default urban speed limit, replacing the previous 50 km/h default in built-up areas. Research from Sweden, Netherlands, and UK shows this single change reduces urban pedestrian fatalities by 30–50% — because it shifts the majority of pedestrian collisions from the 25% survival band to the 90% survival band.
| VRU Type | Primary Risk Scenario | Key Number | Driver Action |
|---|---|---|---|
| Pedestrian (adult) | Night crossing, dark clothing | Detection: 25–30m (dark clothing, dipped) | Slow below stopping distance before pedestrian zones |
| Child pedestrian | Between parked cars, school zones | Appears at 5–8m — unavoidable at 50 km/h | 20 km/h in residential/school zones |
| Older pedestrian | Mid-crossing when signal changes | 0.8–1.0 m/s crossing speed | Never proceed until crossing complete |
| Cyclist | Left hook, dooring, rear-end | 1.5m minimum lateral clearance (Irish law) | Dutch Reach when opening door; wait to overtake |
| Motorcyclist | SMIDSY at junction | Motorcycle looks 10–20% slower than it is | Look twice; double speed estimate; wait for larger gap |
| E-scooter | Cycle lane crossings, night visibility | 25 km/h maximum legal speed | Check cycle lane before turning; treat like cyclist |
| All VRUs at night | Overdriving headlights | Dark clothing: detectable at 25–30m only | Reduce speed to within stopping distance of your detection range |
MIT AVT Consortium: avt.mit.edu | MIT CSAIL: csail.mit.edu | MIT AgeLab: agelab.mit.edu | WHO Pedestrian Safety: who.int | RSA Ireland: rsa.ie | Euro NCAP: euroncap.com | ETSC: etsc.eu
In road law, ethics, and Safe System design, the most vulnerable have the highest priority. Pedestrians > cyclists > motorcyclists > car occupants. This is not anti-car sentiment — it is a recognition that those with less protection need more consideration from those with more. A car driver who treats VRUs as inconveniences is a safety risk. A driver who actively protects them is part of the solution.
Close with the survival probability table from Slide 14. Show participants the 10 km/h difference between 90% and 25% survival. Ask: "How many km/h over 30 do you travel in urban areas?" Then: "What is that worth — to you, and to the person who might step out?" This connects the statistics to a personal decision in a concrete and memorable way.