Method and apparatus for providing accurate localization for an industrial vehicle

• US 8,655,588 B2
• Filed: 11/18/2011
• Issued: 02/18/2014
• Est. Priority Date: 05/26/2011
• Status: Active Grant

• Current Assignee: Crown Equipment Corporation
• Original Assignee: Crown Equipment Corporation
• Inventors: Wong, Lisa, Graham, Andrew Evan, Goode, Christopher W.
• Primary Examiner(s): Cheung, Mary
• Assistant Examiner(s): MAZZARA, ANNE MARIE
• Application Number:
  US13/300,041
• Publication Number:
  US 20120303176A1
• Time in Patent Office: 823 Days
• Field of Search: 701/500
• US Class Current: 701/500
• CPC Class Codes: G01C 21/165 combined with non-inertial navigation instruments
• CPC Class Codes: G01S 17/86 Combinations of lidar systems with systems other than lidar, radar or sonar, e.g. with direction finders
• CPC Class Codes: G01S 17/89 for mapping or imaging

Abstract

A method and apparatus for providing accurate localization for an industrial vehicle is described; including processing at least one sensor input message from a plurality of sensor devices, wherein the at least one sensor input message includes information regarding observed environmental features; determining position measurements associated with the industrial vehicle in response to at least one sensor input message, wherein the plurality of sensor devices comprises a two-dimensional laser scanner, and at least one other sensor device selected from an odometer, an ultrasonic sensor, a compass, an accelerometer, a gyroscope, an inertial measurement unit, or an imaging sensor; and updating a vehicle state using the position measurements.

Claims

1. A method of operating an industrial vehicle in a physical environment, wherein:

the industrial vehicle comprises a mobile computer and a plurality of sensor devices;

the plurality of sensor devices comprise a wheel encoder, an IMU, or both, and one or more two-dimensional laser scanners;

the wheel encoder, IMU, or both, provide odometry data of the industrial vehicle;

the two-dimensional laser scanner provides details of the physical environment;

the mobile computer comprises an EBN module that employs a priority queue that receives input messages from the plurality of sensor devices and associates each input message with a data source and an acquisition time stamp;

the plurality of sensor devices have different sampling periods and different sampling delays so that an order in which sensor data from the plurality of sensor devices is acquired is not the same as an order in which the sensor data becomes available to the EBN module;

the industrial vehicle is moved along a vehicle path by utilizing an Extended Kalman Filter of the mobile computer to model the position of the industrial vehicle in a two-dimensional plane as a probability density, use the odometry data to update a predicted position of the industrial vehicle, and correct for error in the predicted position of the industrial vehicle using environmental features extracted from the two-dimensional laser scanner by comparing the extracted environmental features with a known map of the physical environment;

the predicted vehicle position update by the Extended Kalman Filter is delayed until a trigger message initiating the vehicle position update is received by the EBN module; and

the EBN module processes the input messages in the priority queue in the order of acquisition time upon availability of the trigger message.

2. A method as claimed in claim 1 wherein the trigger message is generated when a dead reckoning error associated with the odometry data exceeds a pre-defined threshold.

3. A method as claimed in claim 1 wherein the trigger message is generated when the priority queue exceeds a certain length.

4. A method as claimed in claim 1 wherein the predicted vehicle position update by the Extended Kalman Filter is delayed an amount of time that is sufficient to ensure that none of the input messages are processed out of order of acquisition time.

5. A method as claimed in claim 1 wherein:

one of the plurality of sensors devices has a longest sampling delay; and

the predicted vehicle position update by the Extended Kalman Filter is delayed until an input message is received from the sensor device having the longest sampling delay.

6. A method as claimed in claim 1 wherein the EBN module deletes one or more of the input messages from the priority queue when a current vehicle position estimate has a high confidence.

7. A method as claimed in claim 1 wherein the EBN module deletes one or more of the input messages from the priority queue to reduce resource workloads.

8. A method as claimed in claim 1 wherein the trigger message initiating the vehicle position update is received from one of the two-dimensional laser scanners.

9. A method as claimed in claim 1 wherein successive input messages in the priority queue are:

integrated, used to update vehicle state, and made available for processing upon receipt of a trigger message, if the input message is odometry data;

used to initiate the predicted vehicle position update, if the input message is a trigger message; and

stored in the priority queue with one or more successive input messages without updating the vehicle state or initiating the predicted vehicle position update, if the input message is not odometry data or a trigger message.

10. A method as claimed in claim 1 wherein, prior to processing the input messages in the priority queue, the EBN module rearranges the input messages according to the associated acquisition time stamp using a data source delay associated with each input message.

11. A method as claimed in claim 10 wherein the data source delay comprises an internal system delay associated with a particular sensor device.

12. A method as claimed in claim 10 wherein the data source delay comprises a characteristic measurement delay associated with a particular data source.

13. A method as claimed in claim 1 wherein the known map of the physical environment comprises known environmental features.

14. A method as claimed in claim 1 wherein the known map of the physical environment comprises a list of dynamic environmental features.

15. A method as claimed in claim 1 wherein the EBN updates the predicted position of the industrial vehicle by integrating the odometry data over time.

16. A method of operating an industrial vehicle in a physical environment, wherein:

the industrial vehicle comprises a mobile computer and a plurality of sensor devices;

the plurality of sensor devices comprise a wheel encoder, an IMU, or both, and a plurality of two-dimensional laser scanners;

the wheel encoder, IMU, or both, provide odometry data of the industrial vehicle;

the two-dimensional laser scanners provide details of the physical environment and are mounted at different measurable positions on the industrial vehicle;

the mobile computer comprises an EBN module that transposes laser scan data from the two-dimensional laser scanners to a common reference frame and combines the transposed laser scan data into a single, virtual laser scan; and

the EBN module utilizes the single, virtual laser scan to control movement of the industrial vehicle along a vehicle path.

17. A method of operating an industrial vehicle in a physical environment, wherein:

the industrial vehicle comprises a mobile computer and a plurality of sensor devices;

the plurality of sensor devices comprise a wheel encoder, an IMU, or both, for providing odometry data of the industrial vehicle;

at least one additional sensor device provides details of the physical environment;

the mobile computer comprises an EBN module that employs a priority queue that receives input messages from the plurality of sensor devices and associates each input message with a data source and an acquisition time stamp;

the plurality of sensor devices have different sampling periods and different sampling delays so that an order in which sensor data from the plurality of sensor devices is acquired is not the same as an order in which the sensor data becomes available to the EBN module;

the industrial vehicle is moved along a vehicle path by utilizing an Extended Kalman Filter of the mobile computer to model the position of the industrial vehicle in a two-dimensional plane as a probability density, use the odometry data to update a predicted position of the industrial vehicle, and correct for error in the predicted position of the industrial vehicle using environmental features extracted from the additional sensor device by comparing the extracted environmental features with a known map of the physical environment;

the predicted vehicle position update by the Extended Kalman Filter is delayed until a trigger message initiating the vehicle position update is received by the EBN module; and

the EBN module processes the input messages in the priority queue in the order of acquisition time upon availability of the trigger message.