Learning Safe-by-Design Neural Network Controllers
For safety-critical control applications, this work provides a more adaptive and computationally efficient approach to enforcing safety constraints in neural network controllers.
The paper addresses the conservatism and computational overhead of safety filters for neural network controllers by jointly learning a controller and parameterized control barrier function parameters, eliminating the need for online quadratic program safety filters. The method achieves reliable safety satisfaction with reduced computational cost in simulations.
Safety filters constructed from control barrier functions (CBFs) are commonly appended to pre-trained neural network controllers to enforce safety requirements. However, this decoupled design with hand-tuned, fixed CBF parameters often fails to adapt to the underlying controller, yielding overly conservative solutions. Thus, given a valid CBF, we address these limitations by jointly learning a neural network controller and neural-network-parameterized CBF parameters, enforcing the resulting affine safety constraints by construction and avoiding an online quadratic program (QP) safety filter at run time. To further improve computational efficiency and scalability, we introduce a lightweight projection architecture that enforces constraints without full constraint enumeration. Extensive simulation evaluations demonstrate reliable, scalable safety constraint satisfaction at reduced computational cost.