An obstacle-avoidance model

The outcome

Implement conservative decisions and explain the limits of geometry-only sensing.

An original overview of the information or commissioning sequence.

Verification boundary

Documented target: Ubuntu 24.04 · ROS 2 Jazzy · Gazebo Harmonic where used. Browser labs tested; ROS/Ubuntu/hardware execution not performed here.

Before you begin

Python 3
LiDAR and drive labs
No physical motor connection

Prepare with Read sensors without guessing, Turn wheel speeds into robot motion.

Build, observe, explain.

Define observations

Invalid or absent measurements produce a stop. Save and run this complete function exercise.

obstacle-avoidance-model-1.txt

import math
def forward_speed(ranges, threshold_m=0.35):
    if not ranges or any(not math.isfinite(x) or x <= 0 for x in ranges):
        return 0.0
    return 0.0 if min(ranges) < threshold_m else 0.10
assert forward_speed([0.2, 0.8]) == 0.0
assert forward_speed([0.8, 0.9]) == 0.10
assert forward_speed([]) == 0.0

Expected: All assertions pass.

Study angles

Compare forward rays with all rays in the LiDAR lab.

Expected: A side obstacle is distinct from a forward obstacle.

Plan integration

Document scan QoS, sector, command timeout and velocity interface before simulation integration.

Expected: A clear contract; this function is not a deployed safety controller.

Troubleshooting

Controller used on hardware: Keep this exercise simulation-only; braking/safety validation are absent.
Infinity means free space: Use a defined invalid-data policy.

Take it further

Add noise and braking distance.
Implement a watchdog in a documented simulator.

Documentation and next steps

Explore another project · Return to learning · Report a project issue