Gap measurement for vehicle convoying
A variety of methods, controllers and algorithms are described for identifying the back of a particular vehicle (e.g., a platoon partner) in a set of distance measurement scenes and/or for tracking the back of such a vehicle. The described techniques can be used in conjunction with a variety of different distance measuring technologies including radar, LIDAR, camera based distance measuring units and others. The described approaches are well suited for use in vehicle platooning and/or vehicle convoying systems including tractor-trailer truck platooning applications. In another aspect, technique are described for fusing sensor data obtained from different vehicles for use in the at least partial automatic control of a particular vehicle. The described techniques are well suited for use in conjunction with a variety of different vehicle control applications including platooning, convoying and other connected driving applications including tractor-trailer truck platooning applications.
1. A method of identifying a position of a back of a first vehicle using radar scenes received from a radar unit on a second vehicle, the method comprising:
(a) estimating a position of the first vehicle relative to the second vehicle;
(b) receiving a radar scene sample from the radar unit on the second vehicle, the radar scene including a set of zero or more detected radar object points, each radar object point corresponding to a detected object having a distance from the estimated first vehicle position;
(c) identifying first vehicle radar point candidates within the set of received detected radar object points;
(d) categorizing the first vehicle radar point candidates based on the distance of the corresponding detected objects from the estimated first vehicle position;
(e) repeating steps (a)-(c) a multiplicity of times, whereby the categorized first vehicle radar point candidates include candidates from multiple sequential radar scene samples;
(f) identifying the position of the back of the first vehicle based at least in part of the categorization of the first vehicle radar point candidates; and
(g) determining an effective length of the first vehicle based at least in part on the identified back of the first vehicle.
2. The method as recited in claim 1 further comprising defining a bounding box around the estimated position of the first vehicle that exceeds a maximum expected size of the first vehicle, wherein radar object points within the set of received detected radar object points that are not located within the bounding box are not identified as first vehicle radar point candidates.
3. The method as recited in claim 1 further comprising estimating a speed of the first vehicle relative to the second vehicle, the estimated relative speed having an associated speed uncertainty, wherein radar object points within the set of detected radar object points that correspond to detected objects that are moving at a relative speed that is not within the speed uncertainty of the estimated speed are not considered first vehicle radar point candidates.
4. The method as recited in claim 1 wherein the identified back of the first vehicle or an effective vehicle length that is determined based at least in part on the identified back of the first vehicle is used in the control of the second vehicle.
5. The method as recited in claim 1 wherein steps (a)-(c) are repeated at a sample rate of at least 10 Hertz.
6. The method as recited in claim 1 wherein categorizing the first vehicle radar point candidates includes populating a histogram with the first vehicle radar point candidates, the histogram including a plurality of bins, each bin representing a longitudinal distance range relative to the estimated position of the first vehicle.
7. The method as recited in claim 6 wherein the identification of the back of the first vehicle is only done after the histogram contains at least a predetermined number of first vehicle radar point candidates.
8. The method as recited in claim 6 further comprising comparing properties of the histogram or mean shift clusters derived from the histogram to a known set of data representative of a target partner vehicle to verify whether the first vehicle is the target partner vehicle.
9. The method as recited in claim 6 further comprising applying a clustering algorithm to the first vehicle radar point candidates to identify one or more clusters of first vehicle radar point candidates.
10. The method as recited in claim 9 wherein the clustering algorithm is a modified mean shift algorithm.
11. The method as recited in claim 9 wherein the cluster located closest to the second vehicle is selected to represent the back of the first vehicle.
12. The method as recited in claim 9 wherein the cluster located closest to the second vehicle that includes at least one of a predetermined threshold percentage or a predetermined threshold number of first vehicle radar point candidates is selected to represent the back of the first vehicle.
13. The method as recited in claim 12 wherein the predetermined threshold percentage is at least 10% of first vehicle radar point candidates in the histogram, and the predetermined number of first vehicle radar point candidates is at least 40 or more vehicle radar point candidates.
14. The method as recited in claim 1 wherein Kalman filtering is used in the estimation of the position of the first vehicle.
15. The method as recited in claim 1 wherein the first and second vehicles are tractor-trailer trucks.