Patent 8108092 entitled "Autonomous behaviors for a remote vehicle" states a rather broad method for commanding a retrotraverse behavior for a mobile robot. This concept is far from new and conceived by many in Universities, business, and government labs well before the patent was issued. Despite understanding this concept back in 2001 when I was in school, I looked further back for prior art and found the following SPIE publication.

The SPIE publication from 1996 by Karly Murphy from NIST entitled "GPS Aided Retrotraverse for Unmanned Ground Vehicles" clearly describes the concept of retrotraverse, which appears to be the central claim of this patent. The only difference I see from the approach described in this paper and the patent claims are that the behavior is enacted upon loss of communication with a remote controller. To me, this is inevitable discovery and not patentable if this is the only differentiation. Further, it is pretty ridiculous that such an obvious and extremely general patent was awarded in general.

Below is the text from Section 3 this paper:

Murphy, Karl and S. Legowik. GPS Aided Retrotraverse For Unmanned Ground Vehicles. SPIE Proceedings: Navigation and Control Technologies for Unmanned Systems, Volume 273, May 1996.

  1. RETROTRAVERSE During retrotraverse the mobility controller drives the vehicle, retracing a. previously recorded path. Vehicle position and velocity data provided by the navigation system are used to compute steering, brake, and throttle commands. The taught path is represented as a series of line segments defined by a set of x-y positions, or knot points, that denote the start and end of each path segment. As the scout drives the vehicle during teaching, the navigation data is monitored and a new knot point is recorded every meter. The vehicle's speed is recorded at each knot point and is used to control the speed during playback. The driver indicates a turn point to either the right or left of the path when the vehicle is at an area that is clear of obstacles. When the final position is reached the driver passes control of the vehicle to the remote operator and leaves.

    After the vehicle has completed its ATA mission at the current observation point the remote operator commands a retrotraverse. The mobility controller then starts the engine, shifts to reverse, and begins backing up along the path using pure pursuit steering and velocity control, both described below. As the first turn point is approached the steering wheel is turned, either right or left, steering the vehicle into the clear area. After the vehicle turns 120 deg. it stops and shifts into forward. It then steers back onto the path and continues following the path using pure pursuit steering. A similar procedure is followed at the final turn point, allowing the vehicle to back into the final position. Obstacles such as another vehicle can be detected with a laser range imaging device, described in Section 3.3.


Looking at the face of the patent there were many many references considered by the examiner. From their titles and a quick look at the initial rejection I believe that one or more of them are at least as good a reference as the SPIE article so this ref. is not likely to add anything new.

The image file wrapper in PAIR shows that claim 1 as-filed was initially rejected as obvious and gained allowance by being amended to include the limitation regarding loss of a control signal initiating the reversing and regaining the control signal stopping the reversing sequence. To attack this patent I would look for things having that behavior. As with most patents there are narrower claims that would be harder to successfully attack than claim 1.

"Inevitable discovery" is not a concept in patent law. Everything is eventually inevitable.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.