Sensor fusion

Combining measurements from multiple sensors (cameras, LIDAR, radar, GPS) into a single consistent estimate of the environment for perception and control.

Sensor fusion merges data from different modalities so a system sees one coherent scene instead of conflicting snapshots. A LIDAR point cloud might define obstacle geometry while cameras read lane markings and traffic lights; radar can track moving vehicles through dust or glare. Filters and registration align timestamps and coordinates so planners and controllers act on a unified model.

Autonomous driving stacks at the DARPA Urban Challenge relied on fusion to stay safe in mixed robot and human traffic. Later deep learning pipelines often learn fusion implicitly, but the engineering problem of trustworthy multi-sensor state estimation remains.