This paper proposes a driver assistance function for positioning a vehicle's hitch underneath the center of a trailer's coupler. Due to a rapid development-to-production cycle, the problem was constrained to use basic ECUs, which have limited performance for real-time implementation. Based on this, we developed two practical solutions to automate the hitching process by using efficient methods that do not require any additional high-end sensors, but rely only on the vehicle's existing rear-view camera and other existing sensors.

In this paper we model a wheeled UGV in an elementary manner to determine the effect of obstacle height on major design parameters, such as wheel size, wheelbase, and center of mass height. We consider both static and dynamic modeling approaches and find that consideration of dynamics allows for more freedom in parameter choice.

This paper contains 3 new contributions to our previous work: First, a continuous visual servoing model for the ODIS omni-directional robot is extended to consider rotation about the robot center, instead of the camera center. This is shown to lead to an undesirable limit cycle due to a Hopf bifurcation. Second, 2 new discrete controllers and models for visual servoing ODIS are derived, which are closer to the servoing methods used in experiments. It is shown that the resulting closed-loop systems are locally asymptotically stable. Finally, experiments are presented which demonstrate successful discrete visual servoing.

In this paper it is demonstrated that pronking (or trotting or pacing) is the preferred mode of behavior of a fairly generic quadruped model. This gait often occurs even though no explicit control is used to force it, and thus the system is said to "self-organize." Further, it is demonstrated that, despite the complicated coupled nonlinear dynamics, the system is still amenable to analysis if certain parameters are assumed small, namely the leg masses and the hip spring constants. Explicit conditions are presented for a symmetric pronk. In one condition, the leg length and the hip spring constant determine the pronking height, which has implications for quadruped robot design.

A common problem in a disaster situation, whether man-made or natural, is the need to quickly identify locations of potential victims and potential hazards to human rescuers, with minimal personnel on site at the beginning of the emergency. Robots can be used to supplement available personnel and to assist them in their search. However, it is important that the robots used not pose a danger to rescuers and victims through collisions or causing structural collapse and that they not get in the way of rescue efforts. Another problem in a disaster area is that environmental conditions (e.g., rubble, scrap metal, dust, smoke, and fire) can make it difficult for robots to maneuver, reduce effectiveness of sensors, and hamper communications. We believe the best approach to developing an effective robotic assistant to human rescue teams is through swarms of low-cost, small robots. The robots should be low cost to eliminate concerns about not recovering them after a disaster. They should be small so they don't hinder or pose a risk to humans (rescuers can kick them or step on them if they are in the way and they will be too small to move objects or injure a person through collision). And we believe a swarm should be used to maximize the area covered while maintaining enough contact between robots to maintain a reliable communications link. A swarm of small robots programmed with simple behaviors can achieve these goals while keeping costs down. This paper describes the efforts we have undertaken at Utah State University in developing urban search and rescue robot swarms for the American Association for Artificial Intelligence and RoboCup Rescue Robot competitions. Both of these competitions make use of the standard urban search and rescue test course developed by the National Institute of Standards and Technology (NIST). Our initial efforts to try to determine the bare minimum competence required for a rescue swarm are described, as are our efforts to improve communications between robots and a base station. Our current efforts to implement a flocking behavior in a robot swarm are described, as is the most recent modular, behavioral robot design we are working on. The development of a graphical user interface for use by rescuers to interface with the swarm is also described. Finally, we discuss the direction we think this research needs to take in the future.

ODIS is an omni-directional mobile robot designed to autonomously or semi-autonomously inspect automobiles in a parking lot. Periodically, its position and orientation references need to be reset. This paper considers visual servoing to parking lot lines as one possible approach. Analysis and simulations demonstrate that a surprisingly simple proportional controller in the image coordinates can accomplish position and orientation alignment with parking lot lines. Unlike previous work, no image Jacobian matrix is necessary. Knowledge of the camera focal length is not required, but the camera and vehicle axes are assumed to be aligned, and the vehicle is assumed to rotate about the camera frame's y-axis.

We consider the dynamics of forward motion for a one-legged hopping robot. We analyze the passive dynamics of the leg and find the necessary initial conditions which produce forward hopping. An approximate Poincare return map is derived using perturbation theory, and a control system designed to consume a small amount of energy is applied. Numerical simulations indicate that a moderate range of stable forward speeds may be achieved. The control strategy applied allows a steady-state hopping motion to be maintained using low-bandwidth sensory feedback while introducing little energy into the system.

We compare two previous approaches for hopping height control to the new scheme proposed in this paper. This new approach is an example of work in the developing area of neural-mechanical systems and has some very simplified versions of building blocks observed in nature, including a central pattern generator. Explicit formulas for hopping height and conditions for stability were obtained for all three approaches based on approximate Poincare return maps (not included). We also present a novel robot leg design and experimental data which supports our analysis. Our adaptive periodic forcing approach is shown to be comparable or out-perform the other two methods in terms of bandwidth requirement, hopping height, and stability properties.

The Acrobot is a simple underactuated system consisting of a double pendulum with an actuator at the second joint only. We derive a set of exact trajectories of the nonlinear equations of motion, which involve inverted periodic motions. The trajectories can be made arbitrarily fast by an appropriate choice of the Acrobot mass and length parameters. Next, we present a nonlinear control law and show how it can be applied to the Acrobot to track these trajectories. In simulations we compare tracking results for our controller and one based on pseudo-linearization. The pseudo-linearizing controller produces significant error for a 1 Hz trajectory, while ours produces none. Finally, we present experimental results which demonstrate that the assumptions of the theory were not overly restrictive. In particular, peak-to-peak oscillations of joints as large as 850 were obtained, despite real-world effects, such as joint friction, inexact parameter values, and noisy and delayed joint velocity data.

We consider a new model for human (or biped robot) locomotion, consisting of a spoked rimless wheel of finite width rolling down a slope. In our model, consecutive spokes are on alternate sides of the wheel, and this models the finite leg separation in humans or robots. Full 3D motion is considered, in contrast to McGeer's 2D model. Numerical studies indicate that for a given slope, a single steady-state solution exists, and this corresponds to rolling straight down the slope at a particular average speed. Moreover, this equilibrium solution is asymptotically stable.

An example of a planar hopping robot is considered, which has only one actuated joint. Simulations demonstrate that the robot can perform both sliding and hopping gaits, despite the fact that almost all other hopping robots have at least two actuated joints.

This paper describes approximate return maps for the bound and pronk gaits in a simple quadruped model. Although expressions for the exact maps cannot be obtained, perturbation theory allows derivation of explicit expressions for approximate maps if particular parameters are assumed small. Equilibria of the approximate maps are presented as a function of system parameters and compare well with those of the exact maps. The simple model predicts that, for a given set of parameters, the pronk produces the greatest ground clearance, while the bound has the shortest gait period, thus allowing more frequent interaction of the feet with the ground to avoid obstacles or provide acceleration.

This paper investigates the potential benefit of elastic energy storage in the propulsion of a swimming robot. A discussion of the hydrodynamic models of von Karman and Sears (1938) and Lighthill (1969, 1970) is presented and is followed by an analysis of the input power required for cases with and without elastic energy storage. It is concluded that the addition of linear springs to the system provides a means to reduce the energy required of the power supply. A transfer function for the system is also derived for purposes of controlling a swimming robot.

It has been shown that a suitably designed biped will walk passively, i.e., without actuation or control, down a shallow slope. This paper extends the concept from bipedal to quadrupedal locomotion. A simple rimless-wheel model is analyzed first to provide a few basic insights, followed by a more complex model with freely-swinging legs. The gaits found for the quadruped are similar to, but more efficient than, those of the biped. However, for any reasonable choice of physical parameters for the system, quadrupedal walking is unstable.

This paper presents a novel robot leg design. The leg is powered by a DC motor which drives elastic tendons to turn a foot. The tendons store elastic energy just as in nature. Preliminary experimental results show the leg has a well-defined resonance which can be exploited to produce hopping. We also offer a comparison of our leg with the cat hindlimb. Eventually, the leg will be part of a quadrupedal robot.

Underactuated mechanisms provide significant challenges to the control engineer. In this paper a nonlinear tracking controller is applied to the underactuated Acrobot, a double pendulum with an actuator at only the second joint. Tracking of oscillatory trajectories with joint angle magnitude up to 85/spl deg/ was achieved. Applications to one-legged robot locomotion are described.>

Dextrous robot hands need to be able to determine the pose of objects to reliably grasp and manipulate them. The first contacts with an object can be used to provide an initial estimate of this information if the object is constrained to be of a particular class. The authors consider a simple example of exploiting class constraints: finding the axis of an unknown surface of revolution. Three tactile curvature measurements on a surface of revolution with twice-differentiable sweeping rule function are shown to be sufficient for determining the axis except for certain singular configurations. Position and orientation error uncertainties and experimental results are presented for a cylindrical tactile sensor.>

A new example of a hopping robot is considered, consisting simply of two links (the end of one link acts as the foot) joined by an actuated, revolute joint. For the stance phase a nonlinear controller is derived that maintains the balance of the robot and periodically accelerates the center of mass vertically. For large enough oscillations the robot can slide or take off. If flight is achieved, the drift caused by non-zero angular momentum can typically be cancelled by rotating the actuated joint an integral number of times, and the robot can land in the same configuration in which it took off. This is due to the holonomy of a single rotation of the actuated joint. Results of simulations are presented in which the robot achieves both sliding and hopping gaits.>

Sparse tactile sensing can be used to determine the positions and orientations of objects if the objects are restricted to being of a particular class. Consideration is given to a simple example of exploiting class constraints: finding the axis of an unknown surface of revolution. Three tactile curvature measurements on a surface of revolution with sweeping rule in C/sup 2/ are shown to be sufficient for determining the axis except for certain singular configurations. Position and orientation error bounds and preliminary experimental results are presented for a capacitive tactile sensor with two contacts. Optimal angular spacing was seen to be 90 degrees between contacts. Experiments showed that the method worked with a position error of 0.07 in. and an angle error of 7.3 degrees .>