Abstract: The invention is a modular and hierarchically organized set of computer programs which comprise methods for controlling a system of semi-autonomous automatically guided vehicles, such as mobile robots. The methods include control programs which: execute in stationary control computers; communicate between the stationary control computer programs and corresponding programs which execute in mobile control computers aboard the vehicles; operate independently in mobile control computers aboard the vehicles. The invention allows the system executive program to command the mobile vehicles to start, to stop, to transfer material to or from the vehicles, to change batteries in the vehicles, to park the vehicles at specific points in a factory, to move the vehicles from point to point in a factory, and to remove the vehicles from the factory.

Abstract: The invention is a method of static collision avoidance for multiple automatically guided vehicles (AGVs) on bidirectional paths. It defines the allowable travel path for an AGV as a series of path segments through or between possible destination points (called nodes) which are defined for the factory floor. AGVs can move to or through any of these nodes and can arrive in a predetermined order, according to "rules" defined by the method. The invention provides better performance than conventional AGV systems by allowing multiple AGVs to coexist in the same pathways without collision or excessive queueing in systems which use free-roving AGVs having programmable bidirectional paths. This new ability maximizes the degrees of freedom of AGV movement while minimizing collisions and "deadlock.

Abstract: The invention provides a method for communicating the need to move a mobile robot, such as an AGV, from one location (node) to another, on a node-by-node basis, incrementally, among the nodes established in a physical environment such as a factory. The invention also provides the means to generate the commands necessary to accomplish this motion, by providing specific steering and drive information such as angle and speed to the mobile robot or AGV. The invention accomplishes these tasks in response to direct commands from a stationary controller as well as in response to continuous node-by-node position updates provided by an external navigation system, such as a visual navigation system, or an internal dead-reckoning navigation system, or both.

Abstract: The invention is a method of dynamic collision avoidance for multiple automatically guided vehicles (AGVs) on bidirectional paths. It calculates windows, or areas of computer memory, against which it matches the reported positions of AGVs. The invention acts on information from a central data base in which the factory or operating space of the AGVs is mapped and from AGV position updates supplied by an independently operating visual navigation system. The invention dynamically calculates the locations and dimensions of the windows for each position update. Since these position updates occur in real time, many times each second, the invention provides the means for the AGV system's controller to prevent collisions between the AGVs and other objects without minimizing the degrees of freedom of AGV movement.