Sensor fusion

Combining cameras, lidar and radar so no single sensor’s blind spot becomes the system’s blind spot.

Lidar measures distance directly rather than inferring it and works in darkness; radar penetrates weather and measures velocity directly; cameras supply texture and colour. The argument for fusing them is redundancy across different failure modes.

Every operator currently running driverless service at scale uses fusion. The cost objection that favoured cameras alone has weakened considerably as lidar prices fell by more than an order of magnitude.