ADAPTIVE MODEL PREDICTIVE CONTROLLER FOR TRAJECTORY TRACKING AND OBSTACLE AVOIDANCE ON AUTONOMOUS VEHICLE

Ali Leman, Zulkarnain; Mohammad Ariff, Mohd Hatta; Zamzuri, Hairi; Abdul Rahman, Mohd Azizi; Mazlan, Saiful Amri; Bahiuddin, Irfan; Yakub, Fitri · 2022 · Crossref

DOI: 10.11113/jurnalteknologi.v84.13778

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of accurate trajectory tracking and collision-free motion for autonomous vehicles, specifically during emergency obstacle avoidance maneuvers. The authors identify that existing controllers, such as conventional Model Predictive Control (MPC) and geometric Stanley controllers, suffer from limited performance due to fixed control gains determined via linearization at specific speeds. These fixed gains fail to account for speed variations and tire force saturation, which significantly degrade tracking accuracy during extreme maneuvers. To resolve this, the study proposes an adaptive MPC controller that dynamically adjusts weighting gains based on the vehicle’s instantaneous longitudinal speed. The methodology employs a 7-degree-of-freedom (DOF) non-linear vehicle model to represent the system plant, accounting for longitudinal, lateral, and yaw dynamics, as well as tire forces limited by the friction circle concept. The adaptive MPC controller is designed using a linearized version of this model, which is updated based on current vehicle speed. To determine optimal weighting parameters (Q1, Q2, and S) for the MPC cost function, the authors utilize a Particle Swarm Optimization (PSO) algorithm. These optimized gains are scheduled via a look-up table strategy covering speeds from 10 m/s to 25 m/s. The proposed controller is evaluated through numerical simulations involving double-lane change maneuvers with static obstacle avoidance. Performance is compared against a conventional MPC (linearized at 10 m/s with constant gains) and an adaptive Stanley controller across low, middle, and high-speed scenarios. The simulation results demonstrate that the adaptive MPC significantly improves tracking error performance compared to the benchmark controllers, particularly under high-speed conditions. At a speed of 25 m/s, the adaptive MPC achieved a 27.3% improvement in lateral error reduction compared to the conventional MPC and a 42.3% improvement compared to the adaptive Stanley controller. The adaptive mechanism effectively mitigated the effects of tire saturation and speed variations, allowing for smoother and more accurate trajectory tracking during extreme avoidance maneuvers. The significance of this work lies in its contribution to robust autonomous vehicle control systems. By integrating PSO-based gain scheduling into the MPC framework, the proposed controller overcomes the limitations of fixed-gain approaches, enhancing safety and performance during critical driving scenarios. The study validates that adaptive tuning of control parameters is essential for maintaining stability and accuracy in the presence of non-linear vehicle dynamics and varying operational speeds, offering a viable solution for real-world autonomous driving applications.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• steering pattern