Search

As tire forces approach road friction limits, vehicles may always exhibit performance degradation and even instability. The actual capacity of the available road friction imposes coupled limits on a vehicle’s longitudinal and lateral accelerations. In this paper, a varying speed method is proposed to design feasible speeds and accelerations, which ensure that the autonomous vehicle will not reach the tire-road friction limits during traversing a clothoid-based path. With the consideration of uncertain traction losses and vehicle parameters, a second-order super-twisting (ST) based speed control strategy is proposed to track above feasible speeds based on varying speed method, and to stabilize the sliding-mode variable of the speed tracking error with relative degree 1. Meanwhile, a second-order quasi-continuous (QC) based path-following control strategy is proposed to follow a desired transition path, and to stabilize the sliding-mode variable of the composite path-following errors with relative degree 2. On this basis, the proposed controllers have been verified to lead good robustness for tracking the ideal speeds and following the desired paths. As compared with the boundary of the autonomous vehicle running at a constant speed, the feasible speed boundary using varying speed method without exceeding the tire-road friction limits can be enlarged up to about 1.59 times, which is decided by the direction change between the entry and exit points of the desired path. [ABSTRACT FROM AUTHOR]

Off-road vehicles always experience serious uncertain longitudinal and lateral slips when running on soft and slope terrains, and some parameters of the vehicles, such as the cornering stiffness and the slip of wheels, are always not constants. In this paper, control strategies for the torque of each wheel and the rear-wheel-steering angle are proposed to maintain a stable velocity and approach an ideal reference model for the off-road vehicle by using second-order sliding mode (SOSM) techniques. An observer is constructed to estimate the actual sideslip angle of the vehicle with the consideration of uncertainties and disturbances of the system. Then, with conditions of bounded uncertainties and disturbances, composite super-twisting (ST) controllers combined with a velocity controller are designed to generate the total torque, the differential torque, and the rear-wheel-steering angle. On this basis, the proposed controllers have been verified to lead good robustness for maintaining the stable velocity and approaching the ideal reference model by using an optimal torque allocation controller at a lower layer. In comparison with conventional yaw moment controllers without the rear-wheel-steering control, the proposed controllers are shown to be more effective. [ABSTRACT FROM AUTHOR]

The lunar surface is a complicated and unknown terrain, and hence, astronauts must be trained to travel on that surface by using a prototype vehicle on the Earth's terrain. However, the gravity of the Earth is six times that of the Moon; thus, the acceleration of a prototype vehicle on the Earth will differ from that of a manned lunar rover on the Moon, when the same acceleration handling input is provided. To obtain the same acceleration in the prototype vehicle on the Earth as that obtained on the Moon, an imitation handling ratio describing the relationship between the acceleration handling input and the wheel rolling speed is proposed in order to replicate the lunar gravitational conditions on the Earth. First, a whole-vehicle acceleration model was established based on the mechanical model of a deformable wheel with a mesh surface. Thereby an acceleration imitation algorithm was derived. A prototype vehicle for use on the Earth is proposed for the purpose of imitating the manned lunar rover under four different vehicle conditions. Fitted functions were used in order to obtain the imitation handling ratio. Finally, experiments were conducted to verify the imitation method; the results revealed that the prototype vehicle on the Earth could effectively approximate the same accelerations experienced by a manned lunar rover, when equivalent acceleration handling inputs were provided. [ABSTRACT FROM AUTHOR]