Navigation module for a semi-autonomous vehicle

• US 5,436,839 A
• Filed: 10/26/1992
• Issued: 07/25/1995
• Est. Priority Date: 10/26/1992
• Status: Expired due to Term

• Current Assignee: Nevada Asset Liquidators LLC (Intellectual Ventures LLC)
• Original Assignee: Martin Marietta Corporation
• Inventors: Carey, Bernard J., Bonissone, Piero P., Dausch, Mark E.
• Primary Examiner(s): Teska, Kevin J.
• Assistant Examiner(s): ZANELLI, MICHAEL J
• Application Number:
  US07/966,472
• Time in Patent Office: 1,002 Days
• Field of Search: 364/424.02, 364/449, 364/460, 364/461, 318/587, 382/1, 382/19, 382/25, 382/26, 395/3, 395/61, 395/94, 348/113, 348/118, 348/119, 348/135, 348/148
• US Class Current: 701/27
• CPC Class Codes: G05D 1/0217 in accordance with energy consumption, time reduction or distance reduction criteria
• CPC Class Codes: G05D 1/0221 involving a learning process
• CPC Class Codes: G05D 1/0231 using optical position detecting means position-fixing by using electromagnetic waves other than radio waves, e.g. optical position detecting means G01S5/16

Abstract

A fuzzy logic based navigation module for use on a semi-autonomous vehicle computes the safety of moving from predefined local regions to adjacent local regions for the purpose of planning a safe path for the vehicle to take to reach its intended destination. Image understanding software provides the input to the system in the form of values for linguistic variables for each local region which impact the safe movement of the vehicle. Using a five level hierarchial rule base structure, the fuzzy inference process computes intermediate and final safety values which are affected not only by local conditions but also by nearby conditions in order to allow a minimum clearance for the vehicle.

Claims

1. In a navigation system for a semi-autonomous vehicle comprising sensors for sensing conditions on terrain within sensor range and producing raw image data about said terrain, and an image understanding module for extracting features from said raw data in the form of feature data for a plurality of local regions on said terrain, a method for processing said feature data to determine the relative safety of moving said semi-autonomous vehicle from each said local region to an adjacent local region, said method comprising the steps of:

defining linguistic variables for each of said features using fuzzy membership distributions;

defining fuzzy rules which determine a safety value for a given combination of values for said linguistic variables; and

for each local region, inputting values for said linguistic variables to said fuzzy rules and determining an output safety value using a fuzzy inference process.

2. In a navigation system for a semi-autonomous vehicle comprising sensors for sensing conditions on terrain within sensor range and producing raw image data about said terrain, and an image understanding module for extracting features from said raw data in the form of feature data for a plurality of local regions on said terrain, a method for processing said feature data to determine the relative safety of moving said semi-autonomous vehicle from said local region to an adjacent local region, said method comprising the steps of:

(a) defining linguistic variables for each of said features using fuzzy membership distributions;

(b) defining fuzzy rules which determine regional safety values and directional safety values for given combinations of values for said linguistic variables for a plurality of local regions;

(c) inputting values for said linguistic variables for one said local region and adjacent local regions to said fuzzy rules and determining a plurality of directional safety values for said local region; and

(d) repeating step (c) for each local region on said terrain.

3. The method of claim 2 wherein said fuzzy rules are organized into a plurality of levels and wherein the outputs of each lower level are inputs to a higher level.

4. The method of claim 3 wherein regional safety values and directional safety values of a local region affect regional and directional safety values for adjacent local regions at a plurality of levels.

5. In a navigation system for a semi-autonomous vehicle comprising sensors for sensing conditions on terrain within sensor range and producing raw image data about the terrain, and an image understanding module for extracting features from the raw data in the form of feature data for a plurality of local regions on the terrain, a method for processing the feature data for determining the relative safety of moving the vehicle from one local region to an adjacent local region, the method comprising the steps of:

considering the terrain as represented by a cluster of seven hexes having a center hex adjacent to each of six other hexes;

determining a regional safety value for each hex of the cluster in response to respective characteristics of the feature data for each hex of the cluster; and

determining a respective directional safety value for traveling from the center hex to each of the other hexes of the cluster in response to the regional safety value of the center hex and of the corresponding adjacent hex of the cluster.

6. The method as in claim 5, further including determining an updated regional safety value for the center hex in response to the regional safety values of the other hexes of the cluster.

7. The method as in claim 6, further including determining a respective first updated directional safety value for traveling from the center hex to each of the other hexes in the cluster in response to the directional safety values.

8. The method as in claim 7, further including determining a respective second updated directional safety value for traveling from the center hex to each of the other hexes in the cluster in response to the first updated directional safety values and the updated regional safety value of the center hex.

9. The method as in claim 5 wherein the characteristics of the feature data are selected from the group consisting of dust, loose sand, packed sand, loose gravel, packed gravel, rocks, slope of terrain between adjacent hexes and combinations thereof.