NeuraPilot project
We are thrilled to announce the latest developments in our NeuraPilot project. We are excited to share that we have successfully implemented and are currently testing two of the three crucial components of self-driving technology:
- Decision Making: Completed.
- Vehicle Control: Completed.
- Environment Detection: Work-in-progress.
To showcase the remarkable capabilities of NeuraPilot, we have created a captivating first-person view 3D and 360 videos. The simulation features twenty cars, each operating as a separate neural network. Remarkably, these neural networks consist of only fifty neurons, making efficient use of resources. The neural networks are trained from scratch, without any pre-training.
What sets our approach apart and enables our cars to learn quickly and effectively is our utilization of a small set of relevant invariants obtained through careful consideration of symmetries. For further information and details, please refer to our publications.
