Autonomous Systems Simulator

Autonomous Rover

Sensor range 9 m

Obstacle density 14

Planner aggressiveness 0.6

Speed

–

Nearest obstacle

–

Distance to goal

–

Goals reached

–

How it works

The rover casts a fan of LIDAR rays every frame, each one measuring distance to the nearest obstacle along its bearing. Where a ray is blocked close by, the sensor reports danger; where it travels far, that direction is open. A simple reactive planner turns each ray's clearance into a repulsive vector — near hits push harder — and blends the sum of those pushes with a vector pulling toward the goal. The rover steers along the combined vector, so it naturally arcs around clutter without ever building an explicit map.

steer = attract(goal) + Σ w(dᵢ)·repel(rayᵢ)
w(d) = clamp(1 − d/range, 0, 1)²
heading += (atan2(steer) − heading) · turnRate·dt

Sensor range — how far each LIDAR ray can see; a longer range lets the planner react earlier and steer more smoothly.
Obstacle density — number of pillars scattered in the arena; more clutter forces tighter, more frequent course corrections.
Planner aggressiveness — how strongly the repulsive field overrides the pull toward the goal; too low and the rover clips obstacles, too high and it overreacts and oscillates.
LIDAR rays — toggle the visualised sweep to see exactly which beams are reporting close returns (red) versus clear space (green) at any instant.

Real-world relevance: this reactive obstacle-avoidance loop is the innermost safety layer on real autonomous vehicles and warehouse robots — a fast, cheap fallback that runs even when the higher-level path planner or full SLAM map is stale or wrong.

Rover Goal LIDAR — blocked LIDAR — clear

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