NVIDIA DRIVE AGX & Wayve London Autonomous Uber Simulator

Wayve VLA Model:Wayve LINGU-3B

Sensor Stream:30 FPS (360° Cam+Radar)

Action Vectors:12 tokens / horizon

Ego Velocity:18.5 mph

Drive AGX Real-Time Telemetry ● Synchronized

Compute Utilization

68%

Stopping Distance

4.8 m

Trajectory Score

98.4 / 100

Model Latency

14.2 ms

💡 Interactive Sandbox: Click anywhere on the London road to drop dynamic obstacles or drag cyclists to probe Wayve's reactive trajectory recalculations.

DRIVE AGX Tensor Core Budget

Precision & Tensor Quantization INT8 / FP16 Mixed

Camera & Sensor Frame Rate 30 FPS

Target Speed (London Urban Limits) 18.5 mph

Wayve Action Tokens (Real-time Stream)

[T+0.0s] LAT_ACCEL: -0.12 m/s²STEER: +0.03 rad

[T+0.2s] LAT_ACCEL: +0.08 m/s²STEER: -0.01 rad

[T+0.4s] SPEED_TARGET: 18.5 mphCONF: 0.994

[T+0.6s] OCCUPANCY: Clear MarginGAP: 6.4m