Preview prompt
Create a self-driving perception-to-control stack figure. Center a semi-realistic ego car in a bird's-eye-view road scene with a labeled sensor suite: 360 LiDAR, 6 surround cameras, front/corner radar, GNSS/IMU, drawing translucent field-of-view cones outward. Surround it with other vehicles and one pedestrian inside colored detection bounding boxes labeled "car 0.97", "ped 0.91". Below, lay out a six-panel left-to-right pipeline: Sensor Fusion, Perception (3D detection + lane segmentation), Localization & HD-Map, Prediction (trajectory cones), Motion Planning (chosen path among candidates), Control (steering, throttle, brake gauges). Add a square BEV occupancy-grid inset showing free/occupied cells and a small latency bar chart. Include values like 10 Hz pipeline, 200 m LiDAR range, 0.3 s reaction, ego speed 14 m/s. Use clean cool-blue and amber accents, thin connector arrows between stages, and crisp small sans-serif labels throughout.
