Technology
Autonomous Hazard Avoidance Systems for Planetary Landers
Quick fact
A planetary lander's autonomous hazard avoidance system can scan the surface, identify hazards like boulders and steep slopes, and select a safe landing site within about one second—without any human input.
Why this is interesting
Imagine trying to land a spacecraft on a rocky, cratered world millions of kilometers away—where a single wrong choice could end the mission. How does the vehicle pick a safe spot when it only has seconds to decide?
Read the full explanation
Understanding Autonomous Hazard Avoidance Systems for Planetary Landers
When a spacecraft approaches a planet, it might seem logical to pre-plan a landing zone from orbit. But many target areas on Mars or the Moon are too rocky, sloped, or obscured for a safe pre-determined landing. Instead, during descent, the lander uses sensors like cameras and LIDAR to take rapid snapshots of the surface. It processes these images in real time to build a 3D map showing where hazards are: big rocks, steep slopes, or crevasses. Then, a computer algorithm evaluates this map to find the flattest and safest region within reach. The guidance system then steers the lander toward that chosen spot, adjusting its trajectory even at the last moments. This entire loop—sense, detect, decide, and redirect—happens automatically in seconds, because radio delays make live human control impossible.
A deeper explanation
The core mechanism relies on two main processes: perception and decision-making. Perception uses imaging sensors to create a digital terrain model (DTM) of the area below. Stereo cameras provide two perspectives of the ground, letting the computer compute depth and height using parallax. LIDAR (Light Detection and Ranging) sends laser pulses and measures their return time to directly measure distances. These data form a grid of elevation points, but raw elevation data alone doesn't directly identify hazards. The system computes local slopes and roughness—boulders appear as abrupt changes in elevation, and steep slopes manifest as high gradients. The hazard detection algorithm thresholds these values to flag dangerous sites. Any cell in the grid with a slope exceeding a safe limit or a roughness above a threshold is marked as unacceptable. Then, the safe-site selection algorithm scans the map for a continuous region that is large enough to accommodate the lander's footprint and free of hazards, picking the safest candidate often by comparing with terrain-relative navigation data that knows where the spacecraft is. Finally, the guidance system commands the thrusters to move the spacecraft horizontally to the chosen site, continuously updating because new images may reveal hazards that were previously hidden or because the vehicle drifts. This closed-loop operation (sense–decide–act–repeat) is what makes autonomous hazard avoidance reliable, and it's crucial for exploring challenging terrains like Mars' Jezero crater or the Moon's south pole, where steep slopes, rocks, and shadows make non-autonomous landing nearly impossible.