M. Stan, W. Burleson, C. Connolly, and R. Grupen. Analog vlsi for robot path planning. Journal of VLSI Signal Processing, 8(1):61--73, Aug 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....paths to the goal if such solutions are possible. The planning algorithm is robust in the presence of environment model errors (such as un modelled obstacles) and provides a framework for incrementally acquiring environment models (see section 3) VLSI prototypes of the planner have been designed [Stan94] for generating path plans in real time. Thus, paths can be planned in realtime, within safe regions and at both the mission and execution levels. In addition alternate paths around un modelled obstacles can be dynamically computed whenever visual sensors detect such obstacles. Visual stealth ....

....of behaviors, ranging from paths that graze obstacles to paths that are repelled away from them. Rapid Computation: Harmonic function computations may be computed using resistive networks. A VLSI implementation of this scheme is in the prototype stage and analog implementations are discussed in [Stan94]. It is anticipated that grids of modest size (100 by 100) will be able to compute harmonic functions within 100 microseconds [Connolly93] Figure 1: Planning a stealth path from start (S) to a goal (G) on a terrain reconstructed from the DemoB site. 3.1 Experiments The terrain over which both ....

Stan, M. R., Burleson, W. P., Connolly, C. I. and Grupen, R. A., "Analog VLSI for Robot Path Planning", J. VLSI Signal Processing, 8(1):61-73, 1994.

....procedure. The obstacle and goal locations are mapped to boundary conditions for a simple partial differential equation (Laplace s equation) The iterative calculation is simple enough that a massively parallel implementation is feasible [46] Analog implementations have also been described [47, 46, 49]. Connolly has also investigated the relationship between biological motor control and harmonic function based motion planning [12, 14, 11] While harmonic functions are easily computed and are free of local minima they are not without problems. One significant problem is that the gradients can be ....

Mircea R. Stan and Waynge P. Burleson. Analog VLSI for robot path-planning. In Asilomar Conference on Computers, Signals and Systems, pages 915--919, 1992.

....equations and can be computed via a simple iterative procedure. The obstacle and goal locations are mapped to boundary conditions for a simple partial differential equation (Laplace s equation) The iterative calculation is simple enough that a massively parallel implementation is feasible [46]. Analog implementations have also been described [47, 46, 49] Connolly has also investigated the relationship between biological motor control and harmonic function based motion planning [12, 14, 11] While harmonic functions are easily computed and are free of local minima they are not without ....

....procedure. The obstacle and goal locations are mapped to boundary conditions for a simple partial differential equation (Laplace s equation) The iterative calculation is simple enough that a massively parallel implementation is feasible [46] Analog implementations have also been described [47, 46, 49]. Connolly has also investigated the relationship between biological motor control and harmonic function based motion planning [12, 14, 11] While harmonic functions are easily computed and are free of local minima they are not without problems. One significant problem is that the gradients can be ....

M. R. Stan, W. P. Burleson, C. I. Connolly, and R. A. Grupen. Analog VLSI for robot path planning. Journal of VLSI Signal Processing, 8(1):61--73, 1994.

....minima [8] For the control of a physical system this implies that gradient descent on the potential field achieves the goal if a collision free path exists. Formally, this approach minimizes collision probabilities for the robot system [7] and can be realized in efficient VLSI architectures [17]. Given a model of obstacles in the workspace and the kinematic structure of the robot, it yields robust and reactive motion controllers. 4.2 Contact Configuration Control The contact configuration controller uses a contextdependent composition of two other controllers. Since our goal here is to ....

M. Stan, W. Burleson, C. Connolly, and R. Grupen. Analog vlsi for robot path planning. Journal of VLSI Signal Processing, 8(1):61--73, Aug 1994.

.... descent of the harmonic potential minimizes the probability of collisions (Connolly, 1994) and responsiveness there is the potential for analog or parallel digital implementations that are capable of forming motion plans very quickly (McCann and Wilts, 1949; Tarassenko and Blake, 1991b; Stan et al. 1994). Harmonic potentials can be computed over arbitrary, discretized environments such as our configuration space grid, by very fast relaxation techniques. From a biological standpoint, harmonic potentials are useful since they serve as mathematical models for a variety of natural phenomena, ....

Stan, M. R., Burleson, W. P., Connolly, C. I., and Grupen, R. A. (1994). Analog vlsi for robot path planning. Journal of VLSI Signal Processing, 8(1):61--73.

.... descent of the harmonic potential minimizes the probability of encountering a known constraint before achieving the goal[7] and responsiveness the steady state voltage distribution in a resistive array is described mathematically by Laplace s equation suggesting the potential for fast analog [29, 38, 36] or massively parallel digital implementations. In contrast, other techniques [3, 5] require substantial off line computation which prohibits the system from reacting well to unexpected changes in the environment. Harmonic potentials can be computed over arbitrary, discretized environments by very ....

Mircea R. Stan, Wayne P. Burleson, Christopher I. Connolly, and Roderic A. Grupen. Analog vlsi for robot path planning. Journal of VLSI Signal Processing, 8(1):61--73, 1994.

....by Koditschek, and has independently been proposed as a robot path planning mechanism [5] Harmonic potentials offer two chief advantages: 1. They obey the min max property, and thus exhibit no local extrema away from goal and obstacles, 2. They can be computed using resistive networks [6, 7]. Using the latter approach, these potentials can be computed over arbitrary environments, if they are discretized. Computation is extremely fast in this case. Even on a digital computer, relaxation times for computing potentials in the cases shown here were in the millisecond range. Detailed ....

Mircea R. Stan, Wayne P. Burleson, Christopher I. Connolly, and Roderic A. Grupen. Analog vlsi for robot path planning. Journal of VLSI Signal Processing, 8(1):61--73, 1994.

No context found.

Stan M and Burleson W (1992). Analog vlsi for robot path planning. In Proceedings of the Asilomar Conference on Signals, Systems and Computers, 1992. Submitted. 21

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC