Fusing sensors for the AUTOMATED-DRIVING FUTURE

Complex processing architectures, faster networks and even more sophisticated software are vital for 100% accurate Level 4 and 5 systems.

Piloted and autonomous driving systems depend on inputs from many types of sensors arrayed to create a highly accurate view of vehicle surroundings. Fusing all these inputs into a single image of each car, pedestrian and roadway marking takes a sophisticated architecture with hefty amounts of computing power and deterministic communications.

At highway speeds, combining inputs from multiple camera views is a complex challenge by itself. But different sensor types are needed to avoid errors, so vehicle controllers must blend inputs coming from multiple sensors and multiple modalities.