Autonomous Vehicle Safety Distance Calculator

Estimate autonomous-vehicle following distance using vehicle speed, system latency, braking performance, perception uncertainty and a configurable safety margin.

Limitations & Assumptions
Important engineering note: This calculator provides an engineering estimate and educational reference. It is not a substitute for vehicle-specific safety validation, hazard analysis, regulatory approval, or a formal safety case. Actual required following distance depends on the vehicle dynamics, actuator response, road conditions, ODD, sensor performance, control architecture and applicable regulatory requirements.

UDHY Tools / AV Safety Distance Calculator

⚙️ Input Parameters

Ego Speed (km/h)

Lead Speed (km/h)

Ego Decel (m/s²)

AV: 4–8 m/s²

Lead Emerg. Decel (m/s²)

Worst-case: 8–10

System Latency (ms)

Full pipeline

Perception Uncertainty (m)

Safety Margin

📊 Results

🚗

Enter your system parameters to calculate an engineering estimate of the required AV following distance.

Minimum Safe Following Distance

—metres

Latency travel distance—

Braking distance differential—

Perception uncertainty—

Safety margin applied—

Time headway equivalent—

Why AV Safety Distance Matters

What This Calculates

The minimum safe gap between your AV and the vehicle ahead, combining three independent components: travel during system latency, braking distance differential between ego and lead vehicle, and position uncertainty from your perception system. All three must be summed — missing any one produces an unsafe result.

When to Use It

Estimating AV following distance during system design
Evaluating the impact of perception and planning latency
Comparing following-distance requirements across different speeds
Assessing braking-performance assumptions
Exploring the effect of perception uncertainty and safety margins
Supporting engineering analysis and safety-case development

This calculator is an engineering analysis tool. Regulatory submissions and formal safety cases require vehicle-specific evidence, testing, validation and applicable regulatory requirements.

Formula

d_latency = v_ego × t_system d_brake = max(0, v_ego²/2a_ego − v_lead²/2a_lead) d_min = (d_lat + d_brake + σ) × safety_factor headway = d_min / v_ego Reference time headway: 1.5s

Practitioner note (Dr. Dilip Limbu, Moovita): The most underestimated variable is system latency. Engineers often account for perception inference time but forget the CAN bus delay to the brake actuator (typically 15–25ms additional) and the actuator’s physical response lag. Always use measured end-to-end latency from physical testing—not just inference time—when performing vehicle-specific safety analysis or validation.

Built by Dr. Dilip Kumar Limbu · Co-Founder, MooVita · Former Principal Scientist, A*STAR · UDHY Engineering Tools

FAQs on AV Safety Distance

1. What is the minimum safe following distance for an autonomous vehicle?

There is no single fixed answer—the minimum safe following distance for an autonomous vehicle depends on vehicle speed, end-to-end system latency, braking capability, lead-vehicle deceleration, perception uncertainty, and the selected safety margin.
For example, at 60 km/h, with 150 ms of system latency and 6 m/s² deceleration, the calculated following distance will vary depending on the lead vehicle’s braking assumption, perception uncertainty, and safety margin. Under the example assumptions used by this calculator, the resulting distance is approximately 18–25 metres.
Use the calculator below to calculate the following distance for your specific AV system parameters.

2. How does system latency affect AV safety distance?

System latency is the time between an event occurring (e.g. the lead vehicle braking) and your AV's actuators physically responding. During this window, your vehicle continues at full speed. At 60 km/h with 150ms total pipeline latency, your vehicle travels 2.5 metres before any braking begins — this becomes part of the required following distance. Reducing pipeline latency directly reduces required headway, which is why perception and planning compute time matter for safety, not just accuracy.

3. What does ISO 22737 specify for AV following distance?

ISO 22737:2021 is an international standard covering low-speed automated driving (LSAD) systems operating on predefined routes. It specifies requirements related to the operational design domain, system requirements, minimum performance requirements and performance test procedures.
This calculator should not be interpreted as determining ISO 22737 compliance. The time-headway value used in this calculator is an engineering reference parameter and can be configured as part of the calculation. Applicable safety requirements should be assessed against the specific vehicle, ODD, system architecture, validation evidence and regulatory framework.

4. What is the difference between stopping distance and safety distance?

Stopping distance is the total distance required to bring a vehicle to a complete stop from a given speed, including reaction time and braking distance. Safety distance (following distance) is the minimum gap to maintain from the vehicle ahead to ensure a collision can be avoided if the lead vehicle performs an emergency stop. Safety distance accounts for the worst-case braking differential between the two vehicles, perception uncertainty, and the safety factor required for regulatory compliance — all of which stopping distance alone does not.

5. How does perception uncertainty affect safe following distance?

Perception uncertainty is the positional error in your sensor stack's estimate of the lead vehicle's location. A GPS/camera fusion system with 0.5m position uncertainty means the lead vehicle could actually be 0.5m closer than your system believes. This uncertainty must be added directly to the required following distance. Better sensor fusion (LiDAR map-matching, tight camera-radar-GPS fusion) reduces perception uncertainty and can meaningfully reduce required headway at the same speed.

6. Can an AV safely follow more closely than a human driver?

An AV may potentially maintain a shorter following distance than a human driver when its measured end-to-end response latency, braking capability and perception uncertainty support it. However, there is no single human reaction-time or AV-latency value that applies to every scenario.