On-Line path planning for an articulated vehicle based on Model Predictive Control

Abstract

In this article, a novel on-line path planning algorithm for an articulated vehicle, moving in a partially known and sensory based reconstructed environment, and relying on Model Predictive Control will be presented. The proposed algorithm belongs to the family of bug like path planning algorithms and has the capability to take under consideration the real dynamics of the articulated vehicle. Based on: a) an a priori knowledge of the current and the goal points, and b) a partial sensory based awareness of the surrounding environment, the algorithm is able to tune online the articulated steering angle in order to drive the front and the rear parts of the vehicle from avoiding collision with obstacles, while converging to the goal point. The proposed path planning algorithm is able to produce on-line the next reference waypoint, solving the local and sub-optimal problem, while in the sequel a Model Predictive Controller is being utilized for creating the proper control signal, the rate of the articulated angle based on an error dynamics kinematic model of the vehicle. Multiple simulation results are being presented that prove the efficiency of the suggested scheme.

Planning Collision-Free Trajectories for Reversing Multiply-Articulated Vehicles

“…Kinematic vehicle models are often used for path planning of articulated vehicles [5][6][7][8][9][10][11]. These usually have two inputs: the longitudinal speed of the tractor unit and the steer angle of the front axle.…”

“…A 'Rapidly exploring Random Tree' (RRT) was used for a 3-trailer vehicle [8] with onaxle hitching but the performance was not repeatable as some runs failed to calculate a feasible path. Additional examples include using feasible velocity polygons [9], a least squares approach [10], and model predictive control [11].…”

“…In summary, the examples of path planning for reversing of multiply-articulated vehicles found in the literature all have limitations, such as algorithm runtime and repeatability. Some assume on-axle hitching [5,6,8,11], or that hitching is behind the axle [9,10]. A few have only been demonstrated for vehicles with one trailer [5,7,11].…”

“…Some assume on-axle hitching [5,6,8,11], or that hitching is behind the axle [9,10]. A few have only been demonstrated for vehicles with one trailer [5,7,11]. A summary of approaches found in the literature for path planning for articulated vehicles is shown in Table 2.…”

Effect of kinematic parameters on MPC based on-line motion planning for an articulated vehicle

“…However, one of the contributions of this article is to present an overall and complete motion planning approach for an AV, based on the full kinematics 125 of the vehicle, combine it with a MPC scheme, evaluate the performance of the motion planning algorithm in more realistic scenarios and present an extended analysis of the algorithms sensitivity based on numerous simulation results, while extending and strengthening the results presented in [24]. The sensitivity analysis, which will be presented in the next Section is considered as the most important contribution of this ar-130 ticle since instead of providing a couple of simulation results, it evaluates the presented scheme from an extended point of view based on multiple simulation results and based on significant algorithmic parameters that dramatically effect the performance of the proposed scheme, while providing direct insights and discussions.…”