44 12 2016 12 Journal of South China University of Technology Natural Science Edition Vol 44 No 12 December 2016 1000-565X201612-0074-07 * 100083 ADAMS ADAMS-Matlab /Simulink 10% 90% U461 1 doi10 3969 /j issn 1000-565X 2016 12 011 1 2-5 1 1 1 6 1 f θ = v fsin γ + l rγ 1 l f cos γ + l r v f θ f γ l f l γ P f = x f y f 2016-05-12 * 863 2011AA060404 Foundation itemssupported by the National High-Tech R & D Program of China863 Program2011AA060404 1955- E-mail wmzhang@ ustb edu cn 1975- E-mailyangjue@ ustb edu cn
12 75 2 3 Fig 1 1 Trajectory analysis of articulated vehicle θ f γ x f cosθ f 0 y sinθ f 0 f = sinγ v f + l r γ θ f l f cosγ + l r l f cosγ + l r γ 0 1 1 2 ADAMS 2 35 t 1 Table 1 1 35 t Fig 2 Steady-state steering experiment based on ADAMS model Basic design parameters of 35 ton mine articulated dump truck /t 28 8 /t 63 8 /t 35 /km h - 1 35 /mm 2278 /mm 5120 / 45 mm/ 1011 5 /961 4 7 E 1 12 3 m /s 2 80 s 500 4 Fig 4 Comparison of lateral acceleration of steady state steering Fig 3 2 3 ADAMS Articulated angle input of steady state steering
76 44 Δw j k= ηe c k yk uk h j 6 w j k= w j k - 1+ Δw j k+ αδw j k 7 η η 0 1α α 2 0 1w j j h j j 2 1 3 4 10 Sigmoid 1 4 5 8-9 Ps Fs Bs δ ft f sw e t yt - y p t 7 fs e δ T p T d T h T c C 0 δ * sw δ sw Vs 5 Fig 5 Driver preview tracking model 6 10-11 Fig 7 6 Fig 6 Neural network-based driver preview model y m k 15 km /h ek= y m k- yk 3 8 Ek= 1 b =10m a 1 - a 0 =24m s 2 =30m s 3 =25m s 4 =30m 2 ek2 4 9 uk= h 1 w 1 + h 2 w 2 + + h m w m 5 7 - Driver-vehicle closed loop system model ADAMS 12 2 2
12 77 10 Fig 10 mm Simulation roadway & planunitmm 8 Fig 8 Ideal experience path 761 0 ~ 135 11 CAN Fig 9 9 Simulation results of neural network-based driver model lane change 3 3 1 11 Fig 11 Flow chart of the control of unmanned driving 3 2 10 40 m 190 0 25 190
78 44 12 12 761 Fig 12 Real road and data display 13 x - 270 mm k = x k - x min 8 x max - x min - 7 9 x k x min 280 mm 10 x max 8 11-13 761 5 14 P xi = e - d2 i /10 31 P xi d i 13 Fig 13 Driver's driving data 761 9
12 79 14 8 20 0 2085 0 5231 0 8799 0 9920 0 7498 0 379 8 0 129 0 3 3 14 15 Fig 15 Neural network-based driver driving data 15-54 mm - 0 47 15 4 14 Fig 14 Head back data - 737 mm - 21 760 mm - 54 mm - 0 47 22 14 10% 200 mm 90% 3 s 100 mm 1 J 2003 31921-23 LIU Jin-xia ZHANG Wen-ming Dong Cui-yan Articulated dump truck and rigid self unloading J Mining Ma- chinery 2003 319 21-23 15 2JOHN Chadwick Advanced mine automation J Mining
80 44 Magazine 1996 1755258-263 3TASCILLO AMILLER R An in-vehicle virtual driving assistant using neural networks C Proceedings of the International Joint Conference on S l IEEE 2003 2418-2423 4DEMIRLI K KHOSHNEJAD M Autonomous parallel parking of a car-like mobile robot by a neuro-fuzzy sensorbased controller J Fuzzy Sets and Systems 2009 160 view optimization neural network driver model J Chi- GUO Kong-hui PAN Feng MA Feng-jun et al A pre- 19 2876-2891 nese Journal of mechanical engineering 20031 26-28 5SEBASTIAN Thrum MAREN Bennewitz WOLFRAM Bur- 11 ADAMS gard et al A second-generation museumtour-guide robot D 2008 J IEEE International Conference on Robotics & Auto- 12 Adams Matlab mation 1999 31999-2005 D 6 2009 J 2015 10198-13 203 D 2014 ZHAO Xuan YANG Jue ZHANG Wen-ming et al Agricultural wheeled articulated vehicle sliding mode trajectory tracking control algorithm J The Chinese Society of Agricultural Engineering 201510 198-203 D 2012 9POMERLEAU D Efficient training of artificial neural network for autonomous navigation J Neural Computation 1991 3188-97 10 J 2003126-28 14 J 19941113-18 WANG Wu-hong Automobile driver behavior patterns and their psychological factors on the reliability of the 7 impact of J Automotive Technology 199411 13-18 D 2013 15 8 D 2013 A Neural Network-Based Autonomous Articulated Vehicle System Considering Driver Behavior ZHANG Wen-ming HAN Hong-bing YANG Jue YI Xiao School of Mechanical EngineeringUniversity of Science and Technology BeijingBeijing 100083China AbstractIn view of the steering characteristics of articulated dump trucksan autonomous articulated vehicle system is proposed based on neural networks and by considering driver behaviors Firsta sensor collecting system based on the laser radar and the angular transducer is establishedand an articulated vehicle kinematics model and a dynamics model of articulated dump trucks are constructed by analyzing the steering characteristics of articulated dump trucks Thenby using the ADAMS softwarea dynamic model of the trucks is constructed to perform a steady state test Moreovera driver model of the artificial neural network control algorithm is constructed based on the optimal preview controland it is verified by an Adams-Matlab /Simulink co-simulation Finallythis control model is also verified by establishing a simulation ground tunnel to perform the straight-road-return and curve-roadfollowing tests The results show thatwhen the constructed control model is applied to the variable curvature road the lateral position error is less than 10% of the passable distanceand 90% of the course angle deviation is optimizedwhich indicates that the constructed control model has a high convergence speeda good steady state and an excellent unmanned driving performance Key wordsdriver modelneural networksdynamic modelco-simulation