Obstacle Avoidance in Real Time With Nonlinear Model Predictive Control of Autonomous Vehicles
Abstract:
A Nonlinear model predictive control (NMPC) for trajectory tracking with the obstacle avoidance of autonomous road vehicles traveling at realistic speeds is presented in this paper, with a focus on the performance of those controllers with respect to the look-ahead horizon of the NMPC. Two different methods of obstacle avoidance are compared and then the NMPC is tested in several simulated but realistic tracking scenarios involving static obstacles on constrained roadways. In order to simplify the vehicle dynamics, a bicycle model is used for the prediction of future vehicle states in the NMPC framework. However, a high-fidelity, nonlinear CarSim vehicle model is used to evaluate the vehicle performance and test the controllers in the simulation results. The CPU time is also analyzed to evaluate these schemes for real-time applications. The results show that the NMPC controller provides satisfactory online tracking performance in a realistic scenario at normal road speeds while still satisfying the real-time constraints. In addition, it is shown that the longer prediction horizons allow for better responses of the controllers, which reduce the deviations while avoiding the obstacles, as compared with shorter horizons.
EXISTING SYSTEM:
Numerous efforts have been made to enable the truly autonomous operation of vehicles using MPC. Sutton and Bitmead [5] considered the problem involving a constrained submarine in order to explore the practicalities of the nonlinear model predictive control (NMPC) problem implementation. They used a discrete nonlinear model to represent a submarine operating in continuous time. The NMPC used a gradient descent optimizer to compute the control policy, and an extended Kalman filter for state estimation. Kim et al. [6] investigated the feasibility of nonlinear model predictive tracking control (NMPTC) for autonomous helicopters and there NMPTC algorithm was formulated for planning the paths under input and state constraints, and was implemented online using a gradient-descent method. Eklund et al. [4] proposed a supervisory controller for pursuit and evasion of two fixed-wing autonomous aircraft. NMPTC was used for real-time trajectory generation and tracking of the evader (which included tracking the pursuer in order to avoid it) as well as the pursuer aircraft (which included tracking the evader in order to target it). This NMPTC controller was then integrated in an unmanned aerial vehicle, which participated in pursuit-evasion games against a U.S. Air Force pilot-operated F-15 aircraft. Fahimi [7] designed an NMPC law for controlling multiple autonomous surface vessels in arbitrary formations within environments containing obstacles. Vougioukas and Ampatzidis [8] used NMPTC for the precision guidance of agricultural tractors to assist farmers in their day to day activities, as well as more generally for four-wheeled vehicles [9]. Presently, many researchers are focusing on active steering methods to reduce the number of motor vehicle accidents, thereby making the driving experience safer [10], [11].
PROPOSED SYSTEM:
This paper presents a Nonlinear model predictive control, which is used to autonomously steer a vehicle around previously unknown obstacles in real time at realistic urban road speeds. This paper builds on earlier work, which demonstrated the feasibility of NMPC in real-time control [13] in terms of complexity and the types of obstacles that can be handled by the controller. Obstacles are included as soft constraints using a pointwise repulsive potential function. As well, a four-wheeled vehicle is used, which has more complex kinematic constraints than the omnidirectional robot used in other previous research [14], [15]. Four-wheeled vehicles also provide a more common platform for practical applications. Two different methods of defining the obstacle are evaluated and it is shown that for larger obstacles the method, which is used, will significantly affect the outcome. The controller is then tested in a realistic simulated scenario of driving on a two-lane road with unanticipated obstacles and lane change requirements. This is done for various vehicles speeds and with various horizon lengths, illustrating different properties of NMPC control. An analysis of the computational time that was used by the NMPC algorithm and the resulting realtime properties follows.
CONCLUSION:
In this paper, an MPC-based trajectory controller is presented for obstacle avoidance in an autonomous ground vehicle in realistic road conditions. The NMPC system uses a simplified bicycle model within the controller but is validated by implementing it in a fully nonlinear CarSim model and tested in various obstacle avoidance scenarios. Additional testing was presented which shows that the NMPC method can handle dynamic trajectory changes and unanticipated obstacles at normal road speeds. To do this, the NMPC used look-ahead horizons considerably longer than in related work for ground vehicles. These tests demonstrated that longer horizons enable better obstacle avoidance, although at a computational price. The computational load of this controller is shown to be within an acceptable range for real-time implementation. Future work will include optimizing the implementation for processing speed and it is anticipated that real-time processing speeds will be achieved.
REFERENCES
[1] D. Portugal, L. Marques, and M. Armada, “Deploying field robots for humanitarian demining: Challenges, requirements and research trends,” in Proc. 17th Int. Conf. Climbing Walking Robots Support Technol. Mobile Mach., Poznan, Poland, Sep. 2014, pp. 21–23.
[2] N. Michael et al., “Collaborative mapping of an earthquake-damaged building via ground and aerial robots,” J. Field Robot., vol. 29, no. 5, pp. 832–841, 2012.
[3] S. Whitsitt and J. Sprinkle, “A hybrid controller for autonomous vehicle lane changing with epsilon dragging,” in Proc. Amer. Control Conf. (ACC), Jun. 2014, pp. 5307–5312.
[4] J. M. Eklund, J. Sprinkle, and S. S. Sastry, “Switched and symmetric pursuit/evasion games using online model predictive control with application to autonomous aircraft,” IEEE Trans. Control Syst. Technol., vol. 20, no. 3, pp. 604–620, May 2012.
[5] G. J. Sutton and R. R. Bitmead, “Performance and computational implementation of nonlinear model predictive control on a submarine,” in Nonlinear Model Predictive Control (Progress in Systems and Control Theory), vol. 26. Basel, Switzerland: Springer-Verlag, 2000, pp. 461–472.
[6] H. J. Kim, D. H. Shim, and S. Sastry, “Nonlinear model predictive tracking control for rotorcraft-based unmanned aerial vehicles,” in Proc. Amer. Control Conf., vol. 5. 2002, pp. 3576–3581.
[7] F. Fahimi, “Non-linear model predictive formation control for groups of autonomous surface vehicles,” Int. J. Control, vol. 80, no. 8, pp. 1248–1259, Oct. 2007.
[8] Y. Vougioukas, N. Sigrimis, K. Arvanitis, and S. Ampatzidis, “A nonlinear model predictive path tracking for precision guidance,” in Proc. Int. Commision Agricult. Eng., Rome, Italy, 2006, p. 1–6.
[9] S. Vougioukas, “Reactive trajectory tracking for mobile robots based on non linear model predictive control,” in Proc. IEEE Int. Conf. Robot. Autom., Rome, Italy, Apr. 2007, pp. 3074–3079.
[10] J. Sprinkle et al., “Model-based design: A report from the trenches of the DARPA urban challenge,” Softw. Syst. Model., vol. 8, no. 4, pp. 551–566, 2009