|
A Novel Steering Control for Real Autonomous Vehicles via PI Adaptive 3 P5 M1 c) t5 I4 \! Y: U
Dynamic Programming
4 f& H- l3 m$ |' u6 \# r
' N. {7 u& w4 q7 q8 W, p+ u, }* p1 R, I5 U* o G/ |
& a& X( n" v: j' J+ h6 Z& Z/ r/ e
. X8 y( K, B5 D5 d
2 A' m0 x: J* L: M; oA novel steering controller for autonomous vehicles is proposed in this paper. For autonomous vehicles, the efficiency 4 Y# n9 V$ r8 S( ~
is an important indicator, in order to reduce the training cost and improve the real-time performance while the car is running on ) h z1 u$ v( q1 j( q
the road, Offline plus online training is proposed in this paper. Offline plus online training is presented to build a model network
4 G _$ ~$ o( }: Gof vehicle steering and correct it in real time. Then we adopt a policy iteration adaptive dynamic programming (ADP) to get the: C3 r9 M4 i0 @# a' h
optimal control law and deploy it to autonomous vehicles. Eventually, the performance of the controller was evaluated using
' D) I9 L1 r# I* Q5 UMatlab simulation. Results verify the feasibility of the theory and show the steering controller is one kind of effective, smooth ; I, G0 r9 I6 Q& |0 O6 R) b
controller for autonomous vehicles' system. + i, l7 @5 R8 |
' A( c' K/ \! Z3 W' d! |2 r5 B: Y& j
|