|
A Novel Steering Control for Real Autonomous Vehicles via PI Adaptive / Z. Z+ M' n- X1 b% D
Dynamic Programming / f5 h/ X/ S# q0 ?
) q& p# ~& q {8 g$ Z: ]- k1 Q% Z! _7 E
* Z! ?: Y+ U% y F2 e- s
+ i& Y1 b+ ?6 W3 ^/ U! u: H) ^. K
6 M$ H% `$ v7 D9 f. K3 w3 P$ D1 K5 ]A novel steering controller for autonomous vehicles is proposed in this paper. For autonomous vehicles, the efficiency 0 n2 k; a8 m! S: j; _" G# m6 m
is an important indicator, in order to reduce the training cost and improve the real-time performance while the car is running on
- W8 q2 L2 H5 U5 g2 Nthe road, Offline plus online training is proposed in this paper. Offline plus online training is presented to build a model network
+ z: L, A( g" i0 n4 g4 Uof vehicle steering and correct it in real time. Then we adopt a policy iteration adaptive dynamic programming (ADP) to get the& ~' S1 A3 ]# M5 {0 T: K
optimal control law and deploy it to autonomous vehicles. Eventually, the performance of the controller was evaluated using " w$ V4 a6 K* q' {
Matlab simulation. Results verify the feasibility of the theory and show the steering controller is one kind of effective, smooth 0 E+ M% j1 _5 E
controller for autonomous vehicles' system. 7 k8 }+ m8 Y% y6 |7 U1 n
0 F7 u8 r0 l# }2 Y3 _4 f) R
4 o! l( g: L3 ~" T; f7 w |