Autonomous Vehicle

Speed 55 km/h

Weather / fog 10%

Traffic density 50%

Sensors

Fusion confidence

Objects tracked

Braking margin

—

Path plan: lane keep.

How it works

Three sensor modalities feed a fusion layer, exactly like a real self-driving stack: LIDAR sweeps a 360° point cloud that is precise but slightly scattered by fog; the camera has the widest semantic field of view but is the most degraded by poor visibility; radar has the narrowest cone yet barely notices weather at all. A planner then checks the lane ahead against the vehicle's own stopping distance and steers around anything it cannot brake for in time.

d_stop = v·t_reaction + v² / (2·a_brake)

confidence = Σ(active_i · w_i · q_i) / Σ(active_i · w_i)
  q_lidar  = 1 − 0.35·fog     w_lidar  = 0.40
  q_camera = 1 − 0.90·fog     w_camera = 0.35
  q_radar  = 1 − 0.10·fog     w_radar  = 0.25

margin = distance_to_obstacle_in_lane − d_stop   (negative ⇒ collision risk)

Speed — the ego vehicle's cruising speed; higher speed means a longer stopping distance and less reaction time.
Weather / fog — degrades camera range sharply, LIDAR mildly, radar almost not at all — the reason real autonomy stacks lean on radar in storms.
Traffic density — how many vehicles and pedestrians share the road, so more objects for the planner to track and avoid.
LIDAR / Camera / Radar toggles — switch a modality off to see fusion confidence and detection drop, just like a sensor dropout in the real world.
Path planning — when a tracked object blocks the current lane inside the braking margin, the car steers to the nearest clear lane; if none is clear it brakes in place (path line turns red).

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.