Neural network-based reference speed self-adjusting brake control algorithm

Abstract

At present, one of the most widely installed brake control algorithms is the Pressure-Bias-Modulated (PBM) algorithm. It performs anti-skid control by pre-setting a reference speed with a fixed deceleration rate. Due to various factors such as wind gusts, road conditions and other impact conditions, the pre-set reference speed cannot effectively adapt to the actual deceleration rate of the aircraft, resulting in lower braking efficiency. Especially for special-purpose airplanes using deceleration parachute, the actual wheel deceleration rate is much greater than the pre-set deceleration rate, which causes the brake control law to misjudge slippage and pressure relief. This paper proposes an algorithm to modify the deceleration rate of PBM reference speed. It uses the change of the equivalent inertia of the aircraft under different landing conditions, and uses the nonlinear mapping of the neural network to make the reference deceleration rate have certain automatic adjustment capabilities, so that the braking efficiency is improved. The simulation model proves that the algorithm is effective.