ANTI-JACKKNIFE CONTROL FOR AUTONOMOUS TRUCK TRAILER PATH FOLLOWING USING PARTICLE SWARM OPTIMIZATION (PSO) AND GENETIC ALGORITHM (GA)

• Main Author: Afghan Fadillah R, M
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/62250
• Institution: Institut Teknologi Bandung

Controlling an autonomous trailer truck system is difficult due to the two rigid parts of the truck (Head and Trailer) and the high probability of a jackknife effect occurring. One of the causes of the jackknife effect is because the speed of the truck is too high or the truck's aggressive behavior when turning. The purpose of this research is to design a control that can avoid the jackknife effect. The method used to avoid this effect is to find the variable value from the tuning results using the constrained optimization method. The optimization methods used are Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) with the objective function being the minimum of Integral Absolute Error (IAE) which has a limit value of maximum steering angle, ?????5?????? ?????????5 and maximum angle between head and trailer, ?????5?????? ????,????? ?????5. A penalty method is proposed to turn a constrained problem into an unconstrained problem. The results showed that by selecting the correct penalty parameter value (?_k), the objective function using the penalty method was able to obtain the value of [???????????? ????????,???????? ????????,?????]????, which caused the system to not have a jackknife effect (it does not exceed the specified limit value). Based on the performance value, that the best result of the value of [???????????? ????????,???????? ????????,?????]????obtained is using the Genetic Algorithm (GA) method with the IAE performance index value of 3688,56, ISE of 3406,79 and ITAE for 2872914