A table-top experiment on formation alignment of three air-levitated robotic vehicles has been performed to demonstrate the feasibility of a more general concept of controlling multiple robotic vehicles to make them move in specified positions and orientations with respect to each other. The original intended application of the concept is in the control of multiple spacecraft flying in formation, as described in “Synchronizing Attitudes and Maneuvers of Multiple Spacecraft” (NPO-20569) on page 64 in this issue of NASA Tech Briefs. In principle, the concept could also be applied on Earth to control formation flying of aircraft or to coordinate the motions of multiple robots, land vehicles, or ships.
The experimental system is, of course, much simpler than a fully developed multiple-robot formation-alignment system would be. In this system (see figure) the three vehicles are levitated over a flat table by air bearings generated from internal supplies of compressed air. Each vehicle is equipped with valves that can be opened momentarily under electrical control to allow jets of compressed air to escape in order to control horizontal translation of the vehicle and/or rotation of the vehicle about a vertical axis.
The problem chosen for the experimental demonstration is to make the three vehicles position and orient themselves at the corners of an equilateral triangle. Even in this simple system, the formation-alignment scheme is highly complex; the most that can be done to describe it in this article is to do so indirectly by summarizing the major features of the equipment and the formation alignment scenario as follows:
The formation-alignment scheme chosen to solve the triangle problem involves the use of lasers, optical sensors, and control subsystems that implement rule-based motion-control algorithms in response to optical-sensor readings. The control subsystems of the vehicles also communicate with each other via radio transceivers.
Each vehicle is equipped with a laser and with an optical-sensor module with an optical axis at an angle of 60° with the laser axis. One of the vehicles is designated the leader, while the others are designated follower 1 and follower 2. The leader initiates the alignment process by activating its laser and rotating to look for follower 1. When the optical sensors of follower 1 detect the laser beam from the leader, follower 1 sends a radio signal to the leader to cause the leader to stop rotating. Follower 1 then performs fine adjustments of its attitude relative to the laser beam from the leader.
Once follower 1 has completed its fine alignment, it turns on its laser and sends a radio signal that commands follower 2 to search for the laser beam from follower 1. This search involves a sequence of prescribed rotations and translations. Follower 2 terminates its search as soon as its optical sensors detect the laser beam from follower 1. Follower 2 then performs fine adjustments of its attitude with respect to the laser beam from follower 1.
Once follower 2 has completed its fine alignment with the laser beam from follower 1, follower 2 turns on its laser and begins a final alignment motion in which it translates along the laser beam from follower 1. When an optical sensor on the leader intercepts the laser beam from follower 2, the leader sends a radio signal that tells follower 2 to stop. At this point, alignment is complete.
This scheme causes the three vehicles to lie at the corners of an equilateral triangle, with the laser of each vehicle aimed at an optical sensor on another vehicle. However, in this scheme, no attempt is made to control the size of the triangle; this is because of the difficulty of optically measuring short distances typical of the intervehicular distances in this experimental system.
This work was done by Fred Y. Hadaegh and Kenneth Lau of Caltech and Paul K. C. Wang and John Yee of the University of California for NASA’s Jet Propulsion Laboratory.