Teacher-Student Knowledge Distillation for Radar Perception on Embedded Accelerators
This work addresses the challenge of real-time radar perception for road safety in automobiles, representing an incremental improvement by adapting existing methods to embedded accelerators.
The paper tackles the problem of inefficient radar signal processing on embedded automotive hardware by proposing a teacher-student knowledge distillation approach, resulting in a student model that runs 100x faster than the teacher model.
Many radar signal processing methodologies are being developed for critical road safety perception tasks. Unfortunately, these signal processing algorithms are often poorly suited to run on embedded hardware accelerators used in automobiles. Conversely, end-to-end machine learning (ML) approaches better exploit the performance gains brought by specialized accelerators. In this paper, we propose a teacher-student knowledge distillation approach for low-level radar perception tasks. We utilize a hybrid model for stationary object detection as a teacher to train an end-to-end ML student model. The student can efficiently harness embedded compute for real-time deployment. We demonstrate that the proposed student model runs at speeds 100x faster than the teacher model.