Patent Application
20130090802
The present disclosure relates to autonomous vehicle following, and in particular the splining of acquired waypoints associated with a leader to define a following path for autonomous vehicle following.
Unmanned ground vehicles capable of autonomous navigation are useful for many purposes and have applications in both military and commercial settings. For example, they may reduce exposure of humans to potentially dangerous situations, reduce manpower requirements in shipping operations, increase traffic flow efficiency and provide pedestrians with hands free operation of a vehicle that can carry their belongings. The state of the art in autonomous navigation generally focuses on the ability to sense and interpret the surrounding environment through on-board sensors and/or a priori map data and make decisions in real-time regarding the path to follow. Such autonomous navigation systems can be complex, expensive and are not always capable of achieving a level of reliability that may be required for a particular application.
What is needed, therefore, is an improved method for autonomous navigation of unmanned ground vehicles that offers increased reliability with reduced cost and complexity.
The present disclosure describes path estimation for autonomous vehicle following based on the acquisition of waypoints associated with a leader. The leader may be a pedestrian or another vehicle which may be manned, tele-operated, or also autonomous with sensors and systems that enable the leader vehicle to derive a correct path to a destination. The following vehicle may acquire waypoints from the leader through use of a global positioning sensor (GPS), a laser incident distance and ranging (LiDAR) sensor or a combination of the two. The acquired waypoints may be pre-processed (including filtering, interpolation and cropping, for example) to improve accuracy and then spline fitted to define a following path that is calculated to estimate the traversed path of the leader. The autonomous following vehicle may then be driven along the calculated path.
The above-mentioned and other features of this disclosure, and the manner of attaining them, will become more apparent and better understood by reference to the following description of embodiments described herein taken in conjunction with the accompanying drawings, wherein:
It may be appreciated that the present disclosure is not limited in its application to the details of construction and the arrangement of components set forth in the following description or illustrated in the drawings. The invention(s) herein may be capable of other embodiments and of being practiced or being carried out in various ways. Also, it may be appreciated that the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting as such may be understood by one of skill in the art.
Throughout the present description, like reference characters may indicate like structure throughout the several views, and such structure need not be separately discussed. Furthermore, any particular feature(s) of a particular exemplary embodiment may be equally applied to any other exemplary embodiment(s) of this specification as suitable. In other words, features between the various exemplary embodiments described herein are interchangeable, and not exclusive.
The present disclosure relates to autonomous vehicle following, and in particular the splining of waypoints associated with a leader and acquired by the following vehicle to estimate a path for autonomous vehicle following. The leader may be a pedestrian or another vehicle, which may be manned, tele-operated, or also autonomous with sensors and systems that enable the leader vehicle to derive a correct path to a destination. The following vehicle may acquire waypoints from the leader through use of a global positioning sensor (GPS), a laser incident distance and ranging (LiDAR) sensor or a combination of the two. The acquired waypoints may be pre-processed (including filtering, interpolation and cropping, for example) to improve accuracy and then spline fitted to a curve that is calculated to estimate the traversed path of the leader as will be explained in greater detail below. The autonomous following vehicle may then be driven along the calculated path.
Providing an unmanned ground vehicle with the capability to identify and follow a leader in an autonomous manner may reduce exposure of humans to potentially dangerous situations, reduce manpower requirements in shipping operations, increase traffic flow efficiency and/or provide pedestrians with hands free operation of a vehicle that can carry their belongings.
Referring now to
Referring now to
Referring now to
As can be seen in this illustration, the acquired waypoints may provide a general indication of the traversed path of the leader 302 but they do not necessarily all lie on a smooth curve. One reason for this is that the acquired waypoints are subject to measurement errors, for example in the GPS or LiDAR systems. For example, waypoint 340 at time 6 is located behind waypoint 318 at time 5 which, if believed to be accurate, would indicate that the path is looping back on itself.
Waypoint 340 may instead be considered to lie outside of an area of interest and thus be removed during a filtering operation on the acquired waypoints to improve the leader path estimation process that results in a following path represented by the curve 350. The location of an acquired waypoint being outside an area of interest may be understood as that situation where the acquired waypoint may lie one to ten degrees (1-10°) outside of a course heading between the immediately preceding waypoint and a subsequent waypoint, including all values in said range in one degree increments. Accordingly, the aforementioned range may be 1-9°, 1-8°, 1-7°, 1-6°, 1-5°, 1-4°, 1-3°, 1-2° or even 1°. As one may further appreciate, the system may be configured to select the particular values depending upon the circumstances of the situation including, for example, the leader speed, the nature of the terrain and the waypoint acquisition sampling rate.
By way of further illustration, waypoint 320 also appears to deviate from the estimated following path 350, to a greater extent than the other waypoints, but the deviation may not be severe enough to place it outside the area of interest and thus, in some embodiments, may survive the filtering operation.
Also, as can be seen in this illustration, adjacent waypoints 312 and 314 are separated by a distance that is greater than the distance between other pairs of waypoints. An interpolation operation may be performed to generate an intermediate waypoint 330 between waypoints 312 and 314, or any other adjacent pairs of waypoints, that are separated by a distance that exceeds a threshold distance. The threshold distance may, for example, be any value in the range of 1 to 500 feet. Accordingly, the threshold distance may be 1 foot, 2 feet, 3 feet, up to 500 feet, in one foot increments. Again, one may appreciate that the system may be configured to select the particular value depending upon the circumstances of the situation. If a single interpolated waypoint is insufficient to fill in a region of sparsely acquired waypoints in this manner, additional interpolated waypoints may be generated. The generation of interpolated waypoints creates a more evenly distributed pattern of points which may reduce the possibility of weighting errors during the curve-fitting process.
Referring now to
Most, but not all, of waypoints 402 behind the following vehicle 102 may be cropped after a period of time as their usefulness in the curve fitting process going forward diminishes. Some of the waypoints 402, however, may be kept to maintain curve connectivity and continuity. For example, when the waypoint 402 behind the following vehicle are computed to generate 5 degrees or more of a course change to the course path 408 (calculated without waypoints 402) the waypoints 402 may be cropped in the ensuing curve fitting process. Here again, one may appreciate that the system may be configured to select the particular values depending upon the circumstances of the situation.
The currently estimated leader traversed path 408 may be further processed to determine the curve length or path distance between the following vehicle 102 and the leader 302. This information may be useful to set or maintain the speed and following distance of the following vehicle 102 so that it does not run out of drivable path.
Referring now to
Acquisition module 502 may be configured to acquire waypoints, associated with the leader traversed path, through wireless receiver 106. The waypoints are transmitted over a wireless communications link between the following vehicle and the leader. The leader may obtain these waypoints from a GPS sensor located with the leader. The GPS sensor may be a high accuracy GPS sensor such as a differential GPS sensor. An alternative, or additional, acquisition module 504 may also be configured to acquire waypoints associated with the leader traversed path. Acquisition module 504 determines the position of the following vehicle from GPS sensor 202 and determines the distance and direction from the following vehicle to the leader from LiDAR 204. GPS sensor 202 may also be a high accuracy GPS sensor such as a differential GPS sensor. The location of the leader may then be determined by translating the GPS based location of the following vehicle by an offset corresponding to the distance and direction to the leader. This determined location of the leader may then be used as the acquired waypoint.
An optional leader identification module 514 may be configured to receive identifying information from one or more leaders and determine the identity of the leader to be followed. Leader identification module 514 may guide acquisition modules 502, 504 to acquire waypoints from the identified leader.
Filter module 506 may be configured to remove waypoints that are located outside of an area of interest. Such waypoints may have been corrupted by measurement errors and their removal may contribute to an improvement in the subsequent curve fitting and path estimation process. The area of interest may be dynamically updated from one waypoint to the next and may generally be defined such that waypoints within the area do not deviate from the estimated leader traversed path by more than an operationally determined threshold amount as previously described. In some embodiments, waypoints that would tend to cause the estimated leader traversed path to loop back on itself would generally be outside of the area of interest. Filter module 506 may also remove older acquired waypoints which lie behind the following vehicle as these waypoints become less useful in the curve fitting process.
Interpolating module 508 may be configured to generate intermediate waypoints in regions of the path where acquired waypoints are sparse, for example spaced from one another by 1 to 500 feet. An interpolation operation may be performed to generate an intermediate waypoint between acquired waypoints that are separated by a distance that exceeds a threshold distance. In some embodiments, the interpolation may be linear interpolation where, for example, a linearly interpolated point (x,y) between a first point (x1, y1) and (x2, y2) may be calculated as:
x=(x1+x2)/2,
x=(y1+y2)/2,
The generation of interpolated waypoints creates a more evenly distributed pattern of points which may improve the curve fitting process.
Curve fitting module 510 may be configured to spline fit the acquired and intermediate waypoints to a curve that is calculated to estimate the leader traversed path using a least-squares fitting process. Spline fitting may group waypoints associated with regions of the path into a number of intervals and then interpolate the waypoints in each of those intervals using polynomials of a selected degree such that each of the polynomial sections fit together in a smooth curve. The coefficients for each polynomial section are calculated to minimize the sum of the squares of the errors (distance) between each waypoint and the resulting curve.
Driving system 512 may be configured to control the propulsion and steering systems of the following vehicle such that the following vehicle drives along the calculated leader traversed path. Driving system 512 may employ a GPS sensor located with the following vehicle to determine the following vehicle's position relative to the path and make navigational corrections as required.
Referring now to
In some embodiments, the waypoints are received from the leader over a wireless communications link, having been generated by a GPS sensor located with the leader. In other embodiments, the waypoints are generated by determining the location of the following vehicle with a GPS sensor, located with the following vehicle, and determining an offset distance and direction to the leader with a LiDAR sensor. The offset distance and direction may then be used to translate the location of the following vehicle into a waypoint associated with the leader. Alternatively, a combination of these two waypoint generation methods may be used.
In some embodiments, previously acquired waypoints may be removed as new waypoints are acquired resulting in an improvement in the curve fitting process. Additionally, the remaining path length along the following path between the current position of the following vehicle and the current position of the leader may be determined and/or updated so that a desired speed and following distance are maintained. This may be useful to prevent the following vehicle from running out of drivable path.
In some embodiments, there may be multiple leaders and the following vehicle may be required to identify a particular leader to follow. The identification may be based on information transmitted from the leaders over the wireless communication link.
The foregoing description of several methods and embodiments has been presented for purposes of illustration. It is not intended to be exhaustive or to limit the claims to the precise steps and/or forms disclosed, and obviously many modifications and variations are possible in light of the above teaching. It is intended that the scope of the invention be defined by the claims appended hereto.
