Autonomous ground care machine field coverage

Abstract

During a staging of an autonomous machine in or near a work region by an operator, a primary orientation for autonomous work is determined. Two or more path definitions are accessed that define different vehicle work paths through the work region and one of the two or more path definitions is selected based on the primary orientation. The selected path definition is used to perform the autonomous work along on the selected work path.

Description

AUTONOMOUS GROUND CARE MACHINE FIELD COVERAGERELATED PATENT DOCUMENTS

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 602,055, filed on 22 November 2023, which is incorporated herein by reference in its entirety.SUMMARY

[0002] The present disclosure is directed to an autonomous ground care machine and system. In one embodiment, a method involves, during a staging of an autonomous machine in or near a work region by an operator, determining a primary orientation for autonomous work. Two or more path definitions are accessed that define different vehicle work paths through the work region and one of the two or more path definitions is selected based on the primary orientation. The selected path definition is used to perform the autonomous work along on the selected work path.

[0003] In another embodiment, a method involves accessing two or more path definitions that define different vehicle work paths through a work region. During a staging of an autonomous machine at an arbitrary location within the work region by an operator, a starting pose of the autonomous machine is determined. One of the two or more path definitions is selected based on the starting pose. The selected path definition defines a selected work path that estimates an intent of the operator during the staging. One or more start points are selected on the selected work path. The selected start point is used for performing autonomous work along on the selected work path.

[0004] These and other features and aspects of various embodiments may be understood in view of the following detailed discussion and accompanying drawings.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The discussion below makes reference to the following figures, wherein the same reference number may be used to identify the similar / same component in multiple figures. The drawings are not necessarily to scale.

[0006] FIG. l is a schematic diagram of a ground care vehicle according to various example embodiments;

[0007] FIGS. 2 and 3 are schematic diagrams of path planning scenarios according to an example embodiment;

[0008] FIGS. 4 and 5 are schematic diagrams of path selection based on operator intent according to an example embodiment;

[0009] FIGS. 6, 7, and 8 are schematic diagrams showing generating a looped work path according to an example embodiment;

[0010] FIGS. 9 and 10 are schematic diagrams showing primary path directions for a work paths according to example embodiments;

[0011] FIGS. 11 and 12 are flow diagrams showing a selection of infill paths according to example embodiments;

[0012] FIG. 13 is a block diagram of a system according to an example embodiment; and

[0013] FIGS. 14 and 15 are flowcharts showing methods according to example embodiments.DETAILED DESCRIPTION

[0014] In the following detailed description of illustrative embodiments, reference is made to the accompanying figures of the drawing which form a part hereof. It is to be understood that other equivalent embodiments, which may not be described and / or illustrated herein, are also contemplated.

[0015] The present disclosure relates generally to work machines such as ground care machines, which may be variously referred to herein as ground care vehicles, ground maintenance machines, ground maintenance vehicles, and the like. Ground care machines,such as lawn and garden machines, are known for performing a variety of tasks. For instance, powered lawn mowers are used by both homeowners and professionals alike to maintain grass areas within a property or yard. The same or different machines may be used for maintenance on the turf areas (and sometimes away from the turf), which may involve performing any combination of operations such as material collection (e.g., plant matter, dirt, golf balls, markers), spraying, fertilizing, dethatching, edging, rolling, towing, snow / ice treatment, snow / ice removal, etc. Other work machines may be used both inside and outside, such as cleaning machines that can be used to clean similar surfaces (e.g., concrete floors) both indoors and outdoors.

[0016] Embodiments of the present disclosure relate to features of ground maintenance machines that facilitate autonomous functionality. Generally, autonomous functionality may include any operation that can be performed without human input that causes or effects a physical action performed by the machine. One example of autonomous operation is autonomous navigation, where the machine can maneuver around a work region without user input, or with minimal user input (e.g., initial placement and initiating a start command). These vehicles are referred to herein, among other things, as autonomous work vehicles, autonomous ground care vehicles, autonomous vehicles, and autonomous machines.

[0017] Some types of autonomous work machines can work using random paths around a work region. So long as the machine has sufficient time to work, it can eventually work the entire region, albeit with some areas being passed multiple times. For some ground care applications such as golf courses, sports playing fields, and the like, a predetermined patterned path is desired and expected, sometimes involving contrasting or overlapping patterns. One example of this is the mowing in crisscross stripes that create a checkerboard effect. For an autonomous mower to create a pattern such as this, a predefined path will typically be created.

[0018] In order to create an aesthetically pleasing mowing path in embodiments described herein, a computer algorithm is used to plan out a work path in a work region. Given a boundary of a work region and descriptions of obstacles (areas to be avoided) within the work region, an algorithm may generate a path that covers the work region insuch a way that certain goals are met, as described for example, in U.S. Patent 1 1,029,691, dated June 8, 2021. The boundaries of obstacles and such are referred to herein as containment boundaries or keep-out boundaries, and may be stored and used together with outer boundaries of work region when defining paths.

[0019] One of the goals of the above-referenced algorithm that the work machine follows a straight line (as defined by the generated path) for as long as possible while covering the field, thereby minimizing slower turns. Other goals are that the generated path avoids defined obstacles within the field and a tool the work machine uses to perform work reaches / covers every part of the field. The generated path geometry will be such that work machine is able to make turns defined by the generated path (e.g., obeying a minimum turning radius of the robot). Other high level performance goals may be used in a path generation algorithm without departing from the scope of this disclosure.

[0020] While automated path generation can be used to effectively guide an autonomous machine through a work region, the ease of operator use is not always considered in path generation. For example, it is common to define a beginning and an end of a generated path, as this is the way the path points may be stored, e.g., as a series of geolocations in a linked list or other data structure with a natural progression from beginning to end. The operator, who performs certain tasks such as staging the work machine and collecting the work machine after work is complete, may not always arrive at the work region at the same location. In some instances, the operator may not even know the location of the predefined starting point.

[0021] A machine that is staged some distance from where the predefined starting point can self-navigate to the starting point, assuming it can dynamically generate a path to the starting point from an arbitrary location. Otherwise, the operator may have to move the machine, e.g., guiding the machine via onboard controls or a remote control. In either case, the movement of the machine to the starting point can take valuable time that could better be used for performing autonomous work.

[0022] Another issue that might occur when staging the work machine relates to the path being directional. Whether the desired path is operator selected or predetermined, the operator may still not know or remember which direction the machine should be placedto follow the path direction. If the machine is placed near the starting point but in the wrong direction, it may need to circle around the same area to achieve the desired orientation to start work. This type of behavior might be seen by the operator as sub- optimal behavior or even a malfunction.

[0023] In embodiments described below, a path planning service for an autonomous machine takes into account operator intent when staging an autonomous machine in a work area. When the operator stages the machine in a work location for example, the machine can begin working via a shortest transition path to the work path. If multiple paths are possible, the orientation of the machine as well as its location may be determinative on which path is taken to perform the work, as well as minimizing transit time needed to start the work.

[0024] In FIG. 1, a block diagram shows details of an apparatus according to one or more embodiments, the apparatus operable as a ground care vehicle. As seen in FIG. 1, the ground care vehicle 100 includes a drive section 102 that moves the ground care vehicle 100 within a work region 104 along a path direction 106. The path direction 106 is generally aligned with a forward or reverse direction of the ground care vehicle 100 when going straight, and tangent to a curved path through a turn. A work unit 108 defines a full coverage width 112 of a path of the ground care vehicle 100. A cross path direction 107 of the full coverage width 112 is defined normal to the path direction. The cross path direction 107 is aligned with a lateral direction of the ground care vehicle 100 as it moves along its path in a forward direction 124.

[0025] As seen in the embodiment of FIG. 1, the work unit 108 is centrally disposed in the cross path direction 107. The work unit 108 is a mechanism that performs a specific operation, e.g., cutting, dethatching, rolling, etc. The work unit 108 may be a reel cutter, rotary cutter, roller, brush, spiker, verticutter, dethatcher, sprayer, etc. The ground care vehicle 100 may have multiple such work units, e.g., arrayed in the cross path direction 107 to cover a greater coverage width.

[0026] The work unit 108 may have a dedicated electric motor, actuator, or the like (not shown), and / or may be driven by a power take off from a main drive unit of the vehicle 100. In either configuration, the work unit 108 may be selectively activated anddeactivated to allow work to stop or start independently of the movement of the ground care vehicle 100. The work unit 108 may have a mechanical and electrical coupling interface to the ground care vehicle 100, e.g., to a central frame 114 and central electrical system 116 of the vehicle 100.

[0027] The ground care vehicle 100 is shown with two wheels 122 used to propel the vehicle. The wheels 122 may be independently driven such that they can differentially steer the vehicle 100. The vehicle 100 may include additional powered wheels, or unpowered wheels such as caster wheels. The drive section 102 may include one or more electrical motors (not shown) that drive the wheels 122, powered by an onboard battery 120. The ground care vehicle 100 may be driven by another source of mechanical power such as an internal combustion engine, although some source of electrical power such as the battery 120 and a generator may be provided to power onboard electrical components.

[0028] The vehicle 100 includes a controller 110 that controls the autonomous and non-autonomous functions of the vehicle 100. The autonomous functions may include sensor collection and fusion, navigation control, work unit control, remote data center communications, automatic stop, etc. The non-autonomous functions may include manual control inputs, user interface indicators / display, manual locks, etc. Various functionality described below related to path planning may be performed entirely off vehicle (e.g., at a data center 118 or a standalone computer), entirely on vehicle (e.g., via controller 110), or cooperatively performed by both the vehicle and a separate computing device or service.

[0029] In FIG. 2, a diagram shows a ground care vehicle 100 starting work in a work region 104 according to an example embodiment. A path definition 202 is accessed that defines a vehicle work path 200 through the work region. The path definition 202 is a set of data that describes the path in a way that it can be used by the ground care vehicle 100 to work within a known tolerance, e.g., defined by locational error of the path definition 202 and the accuracy and precision the vehicle’s own navigation system.

[0030] The path definition 202 may include path geometry expressed in any form, e.g., an ordered collection of points, parametric description of path segments, etc. The path definition 202 may be calculated by the data center 118 and transmitted to the autonomous work vehicle 100, e.g., via a mobile data network, or may be calculated by the machine100 itself. The path definition 202 is generated based on some previously collected reference data, such as geometry of boundary and obstacles of the work region and vehicle specifications (e.g., turning radii, work path width, etc.). Multiple path definitions may be prepared for the work region 104, as will be described in detail elsewhere. The path definitions may also include metadata, such as user-defined name, simplified location data (e.g., bounding box, centroid), last time the path was worked, linkage between associated paths (e.g., paths that can be combined to form a cross hatch or similar pattern), etc.

[0031] In one embodiment, the path definition 202 is created in a training or setup phase based on the user or operator recording a boundary of the work region 104. The boundary can be recorded in a number of ways or receiving a first operator input, such as guiding the work vehicle 104 along the boundary, moving around the boundary with a different measuring device (e.g., GNSS receiver), selecting a path on a graphical map, etc. The user / operator may also define other geometry of the work region 104 in this way, e.g., obstacles, keep out areas, transit paths, etc.

[0032] In one embodiment, at the end of the boundary recording process, the user / operator points the work vehicle in a direction that the user / operator want the first path plan to be oriented, which is a second operator input different from the first. The direction defines an initial angle and can be determined via a number of different second operator inputs, such as selecting a compass heading (e.g., due north), graphical inputs to mobile application, machine orientation in the work region, driving a manual path, etc. Once the initial angle is determined, the first path plan / definition is generated that conforms to the boundary and the initial angle.

[0033] Once the first path definition is generated, the initial angle also defines a predetermined set of angles for path selection at start / staging of work as described below. For example, if no angles are yet determined, then the initial angle is used as the predetermined set, with the ability to add other angles to the predetermined set later. If and when the vehicle is subsequently placed at approximately the initial angle (e.g., within some angle range), the vehicle will begin autonomous work with the assumption that the user intends to use the first path definition. This can be repeated for defining multiple work paths, e.g., after defining the first path, the user reorients the machine and then a second fillpath pl an / defi nition is generated and added to the predetermined set. Both angles of the set are available for later autonomous work and selected in a similar fashion by inferring operator intent during staging. In some embodiments, additional paths can be created later (e.g., after the training phase) and / or based on some other criteria. For example, the user / operator may create, via the user interface, a second path definition with fill lines that are some angle (e.g., 90 degrees) from the first path definition by specifying the angle, e.g., typing in an angle, selecting from a list of angles, rotating a graphic representation. In some cases, such alternatives could be automatically generated in response to a system setting, configuration, or preference.

[0034] As shown, the autonomous work vehicle 100 is staged at an arbitrary location within the work region 104 by an operator. The staging involves, among other things, placing the work vehicle 100 at the location (e.g., manually or via remote control) and initiating autonomous work, e g., via an onboard or remote user interface. During the staging, a starting pose of the autonomous work vehicle 100 is determined, e.g., measured by onboard sensors such as Global Navigation Satellite Systems (GNSS) sensors and a compass. Generally, pose refers to a combination of location and orientation (the latter indicated by start heading 201), and is updated as the work vehicle 100 moves through the work region 104. A pose measurement may include other data, such as roll and tilt angles, but those are not necessarily necessary to be measured in the illustrated scenario.

[0035] A target navigation start region 204 within the work region 104 is determined based on the starting pose. The target navigation start region 204 representing an allowable region within which to start work. As shown, the target navigation start region 204 the target navigation start region 204 encompasses a forward field of view of the autonomous work vehicle 100. This field of view could be defined on the center, forward-most part of the vehicle and / or based on a camera or other sensor at or near that location. The illustrated forward field of view is around 180°, and in other embodiments it could be less, e.g., 120°, 90°, 60° or less. For purposes of this example, the field of view is assumed to be centered along a longitudinal centerline of the vehicle 100, such that a 180° field of view would correspond to ±90° relative to this centerline.

[0036] Note that the field of view that defines the target region 204 is shown as a circular sector, however the target navigation start region 204 need not always conform to this shape. For example, if the circular sector overlaps a boundary of the work region 104 or an obstacle within the work region 104, then the target navigation start region 204 may be truncated to remove these parts from the sector. Also, the field of view angle and extent (corresponding to the circular radius) may be adjusted depending on work conditions, user preference, and the like. For example, if a 60° view at 3 meters fails to result in a viable starting point as described below, then it could be adjusted to 80° and / or 4 meters for a subsequent starting point search.

[0037] Once the target navigation start region 204 is defined, a processor (either on the work vehicle 100 or data center 118) selects one or more start points 206 on the work path within the target navigation start region. The selected start point 206 provides an estimate of an intent of the operator during the staging. Generally, operator’s intent is assumed to be a desire start autonomous work without significant non-work travel and without excessive turning, and is defined at least by the vehicle pose. The operator may provide other indications of intent, such as moving the autonomous vehicle 100 between two points that define an orientation and direction, manipulating a user interface such as a pointer on the vehicle 100 or swiping on a graphical map on a smartphone application. In any of these examples, the user input is enough to infer intent while the user input is not so detailed as to be executed (e.g., via machine instructions) on its own without additional calculations. Note that while the start point 206 is shown on the boundary of the target navigation start region 204 in this example, a starting point may be anywhere within the target navigation start region 204.

[0038] In one embodiment, the determination of the one or more start points on the work path involves considering, in order, nearest points, then angular offset of the points from the start heading of the machine, and then alignment of the machine start heading with a primary direction of the work path. The concept of a “primary direction” will be explained in detail below in the discussion of FIGS. 9 and 10.

[0039] In some embodiments, start points may be less likely to be used (or excluded) if they would result in the autonomous machine 100 making a turn of somemaximum turn angle (e.g., greater than ±90°) to navigate to the start point. This maximum turn angle may be used to select from multiple valid start points, e.g., the start point that requires the least turning of the machine to arrive at is more highly ranked or weighted. The start point selection criteria could also be ranked / weighted by the total distance needed to access the points, e.g., shorter distance has a higher rank or weight. In other embodiments, the selected start point 206 could be a closest point on the path 200 that that can be navigated to via a Dubins path with a radius greater than or equal to a minimum turn radius of the autonomous machine 100. This criterion based on a Dubins path (e.g., the shortest curve that connects two points in the two-dimensional Euclidean plane) could be considered in addition to or instead of the maximum turn angle criterion described above.

[0040] As indicated by dashed line 208, the selected start point 206 is used for performing autonomous work with the autonomous work vehicle 100 along on the vehicle work path 200. The line 208 represents a transit path, and the work implement may be disengaged along the transit path. The start point 206 can also be an end point, in the event that the full work path 200 is worked during the same session. The path in such a configuration is a geometric loop, which is generally a closed curve where the start point is also the end point. Some paths need not be looped, and so in order to allow starting at an arbitrary point, an extra leg of the path may be added in order to complete a loop. This will be discussed in greater detail below.

[0041] The illustrated path 200 includes an infill path 200a (also referred to herein as a fill path) that fills an area of the work region 104 and a clean-up path 200b that follows a boundary of the work region 104. The clean-up path 200b (also sometimes referred to as headland laps conforms to the boundary and may overlap with the infill path 200a, e.g., parts of the infill path 200a where the autonomous machine 100 turns around. One or more transition paths 200c connect the infill path 200a and the clean-up path 200b. The work implement of the autonomous machine 100 may be turned off on the transition path 200c.

[0042] This type of path arrangement lends itself well to be formatted as a loops, because the clean-up path 200b traverses a boundary of the region, and so there may be anumber of locations to place the transition path without undo amounts of redundant navigation, e.g., moving along a path that already has been worked or will later be worked. The path definition 202 may have a beginning point 210 that, for example, is at the beginning of an ordered list. Even with a predefined beginning point 210 in the stored data, the actual traversal of the autonomous machine 100 can start at any point defined within the path definition 202. Because the path 200 is defined as a loop, the beginning point 210 can also be an end point.

[0043] In FIG. 3, a diagram shows a ground care vehicle 100 starting work in the same work region 104 as in FIG. 2, but using a different work path 300 according to an example embodiment. As with the previous example, this path 300 is represented on a computer as a path definition 302, and the path 300 includes a fill path 300a, clean-up path 300b, and transition path 300c. Note that the fill path 300a is a similar parallel line fill as fill path 200a in FIG. 2, but the lines of fill path 300a at a different angle than fill path 200a.

[0044] In FIG. 3, the starting pose of the autonomous machine 100 is the same as shown in FIG. 2, therefore the target navigation start region 204 is the same. Because the path 300 within the target region 204 is different than in FIG. 2, the starting point 306 and transit path 308 is different. The starting point 306 is different than a beginning point 310 that may be defined in the path definition 302 data structure. This path 300 is also a loop, and so beginning point 310 can also be considered an end point. Note that the transit path 308 is longer than path 208 in FIG. 2 and has a smaller curve diameter than path 208 in FIG. 2 due to the larger angle of the autonomous machine 100 relative to the primary direction of the path 300 compared to path 200.

[0045] In some embodiments, the autonomous machine 100 may have multiple paths available from which to choose. Some of those may be selected deterministically, e.g., one of two crisscrossing patterns that are alternatively selected for each subsequent session. The use of alternating directions in such a case may be defined in a work requirement that drives path alignments, e.g., that successive primary path directions between subsequent work sessions vary. In other embodiments, the operator will be able to choose a primary direction of the fill pattern, e.g., due to preference, current turf condition,etc. Therefore, as part of determining operator intent during staging, the data center and / or autonomous machine may select one of two or more path definitions based on the starting pose. The selected path definition defines a selected work path that estimates an intent of the operator during the staging and is used to perform work in the work region.

[0046] In FIGS. 4 and 5, diagrams illustrate how one of two different fill paths 200a, 300a may be selected as a work path based on staging pose. Parts of both paths 200a, 300a are shown using dotted lines within the work regions 104 in both FIGS. 4 and 5. These parts are used to define a primary direction of the entire path. The start heading 201 of the autonomous machine 100 forms different angles 400, 402, 500, 502 with parallel lines of the fill paths 200a, 300a. In FIG. 4, fill path 200a is chosen because angle 400 is less than angle 402. In FIG. 5, fill path 300a is chosen because angle 500 is less than angle 402. Generally, a path definition may be selected by minimizing a difference between a heading of the primary orientation and the primary angle of the selected path definition.

[0047] In this example, just the illustrated parts of the fill paths 200a, 300a were considered as defining a primary direction when comparing to the machine start heading 201, however the entire paths 200, 300 could be considered if the other parts of the paths are within the target navigation start region during staging. Usually, a majority of the work will be done along the fill paths, and so it is likely the fill path directions are of primary concern to the operator, and so in some embodiments only the fill paths will be considered when trying to determine operator intent. In such a case, the fill paths 200a, 300a could be defined within separate path definitions, those definitions being separate from definitions for the clean-up paths 200b, 300b, and transition paths 200c, 300c. For example, a single clean-up path definition could be defined along with multiple fill path definitions and transition path definitions. The final work path in such a case can be obtained by combining multiple path definitions, e.g., final path is a selected one of the fill paths plus the single clean-up path, plus a predetermined or dynamically generated transit path that connects the selected fill path and the clean-up path.

[0048] In some cases, the path selection and generation may be configured such that a path is selected only if primary direction of the path is within some predefined angle of the start heading of the machine, e.g., 30°. If no paths are found that align with the startheading (e g., paths are unacceptable within a threshold angle), then a new path may be generated. The new path may have a primary direction that corresponds to the start heading, or it may “snap” to some predefined value. The use of snap angles may be a primary work requirement, such that the machine will always try to select a path definition with a primary angle set of one angle from a predetermined set of discrete angles.

[0049] For example, the system (e.g., through user preference setting) could autogenerate paths that conform to eight compass directions (e.g., N, NE, E, SE, S, SW, W, and NW) in which case a path is selected if the angle between primary direction of the path and the start heading is less than 22.5° (a random selection could be made if the angle is exactly 22.5°). If no paths satisfy this criterion, then a new path could be defined and used. Generally, the use of “snap” preference, setting, or requirement involves defining a discrete number of primary path angles. Such path angles may be evenly distributed or spaced although this is not required. For example, the increment between N snap angles may be defined as 360 / N and the set of angles starts at M, which is an offset from zero. In the eight-compass-direction example above, N=8 and M=0, thus eight equal angles are 45 degrees from each other, starting at zero: {0, 45, 90, 135, 180, 225, 270, 315}. For a noneven distribution, a non-linear function may define the set of snap angles and / or explicit set of angles may be defined, e.g. { 15, 60, 120, 165, 195, 240, 300, 345}.

[0050] Note that in FIGS. 2 and 3, the paths 200, 300 are already defined as a loop, with beginning points 210, 310 that can serve as end points in the path definitions, even if not used for those functions for autonomous work. It is possible that automatically or manually generated paths may not be in the form of a loop. An example of this is shown in FIG. 6, where a path 600 has starting point 604 on an upper left corner of a rectangular work region 602, traversing horizontally from top to bottom, ending at the bottom right corner at end point 606. The autonomous machine (not shown) turns around in end parts 608, which are often referred to as headlands. The work implement of the machine may be turned off while in the end parts 608, or the work implement can be left on if there is no negative impact on the machine and / or the work.

[0051] As with other embodiments, an operator may want to start within a posebased, target navigation start region which encompasses arbitrary start point 610. In orderto start from the arbitrary point 610 on the path 600, an additional transit path from the end point 606 to the beginning point 604 may be generated. Two such transit paths are shown in FIGS. 7 and 8. In FIG. 7, a transit path 700 remains as much as possible in the side region 608 or close to the border near the top. In FIG. 8, a transit path 800 takes the most direct route from end point 606 to beginning point 604. A work implement may be turned off when the autonomous work machine traverses either of these transit paths 700, 800.

[0052] Note that in a situation such as shown in FIG. 6, where a clean-up or outside border path is not used, a strategy of first maneuvering the autonomous work vehicle to the start point 604 to perform work along path 600 would usually take less time than working loops as shown in FIGS. 7 and 8. Nonetheless, this may be an acceptable tradeoff for the benefit of the user being able to start at any location. Also, the use of the start heading during staging of the machine (or other user input described above) as described in relation to FIGS. 4 and 5 can be used to select between the indicated path 600 and other paths (not shown) that run in different primary directions, e.g., based on a smallest angle between the start heading and the primary direction of the chosen path.

[0053] As described below, the existence of a primary direction of a path may be determined in order to determine operator intent and / or to satisfy a work requirement that is can be set to override apparent operator intent. In cases where a fill path includes mostly parallel lines, the primary direction would correspond to the direction of those lines. There may be other paths where the primary direction may not be so clear, particularly given the ability of autonomous work machines to accurately navigate complex work paths. One example is shown in FIG. 9, where partial work paths 900, 901 for the same work region include parallel curves, e.g., sine waves. In this case, the lines 902, 903 represent the respective primary directions, as they correspond to zero axes of the waves.

[0054] In FIG. 10, partial work paths 1000, 1001 each cover the same two regions with parallel lines in two different directions in each region. This could result in two possible primary directions 1002, 1003 for work path 1000 and two possible primary directions 1004, 1005 for work path 1001. In this scenario, the primary direction may be start-location-dependent, and / or there may be a preference for one of the possible directions.

[0055] In FIG. 11, a diagram shows a method for setting up an autonomous work machine in a work region according to an example embodiment. The method involves recording 1100 outer boundaries and optionally exclusion zones in a work region. This may only need to be performed once in a work region, e.g., initial training of the machine. Thereafter, a stored version of the boundaries (which may include both outer boundaries and inner boundaries of exclusion zones) can be retrieved, e.g., from a database. The other operations shown in blocks 1102-1105 may be performed during an initialization run after recording and thereafter each time the region is worked.

[0056] In block 1102, the operator parks the autonomous work machine (the “unit”) oriented in the desired striping direction. The heading of the unit will provide at least an approximation of the desired striping direction. In some cases, the actual directions will “snap” to a preset number of directions, e.g., eight compass point directions. In block 1103, a line is recorded between two points, e.g., manually navigating the unit from point A to point B, or doing the same with a handheld device that can geolocate such as a mobile device.

[0057] In block 1104, the user can set a striping direction in a user interface (UI). For example, a map view of the work region may be graphically presented on a mobile device touchscreen graphical user interface (GUI), where the user can draw a line, rotate an arrow, etc., to indicate the desired direction. At block 1105, a software component, e g., running locally or on a remote server, can infer the direction and suggest to the user, e.g., via a UI. The inference may be, for example, based on border geometry (e.g., aligned with a major axis of the work region), surrounding infrastructure (e.g., walkways, waterways), previous, manually-produced patterns used in the work region that may have been determined in the past through aerial imaging, etc.

[0058] Following striping determination at blocks 1102-1104, the machine switches 1106 to an auto mode, in which an infdl is generated 1110. Alternatively for block 1104, and also for block 1105, an addition GUI input 1108 may be received in order to edit / confirm aspects of the direction choice, after which the infill is generated 1110. This infill corresponds to a path definition described elsewhere herein, and in someembodiments, the path definition used in autonomous work may be selected from two or more path definitions.[00591 Some aspects related to selecting or generating the infill according to an embodiment is shown in the flowchart of FIG. 12. At block 1200 in, FIG. 12, user intent is determined regarding a striping direction angle (Ostripe), which is a primary orientation for autonomous work. In block 1202, auto mode is entered. In this case, other data 1201 that is used by the auto mode includes at least initial infill No and a striping angle 0Noassociated with No, as well as number of allowed striping directions A. The striping angles described in FIGS. 11 and 12 correspond to the angular orientation of a primary part of the selected path geometry as described above.[0060J In block 1203, the user defined striping direction angle Ostripe is compared to the initial infill striping angle 0NQby a threshold amount a = 907A, where A is the number of allowed striping directions. This is functionally equivalent to determining that an angle between Ostripe (the primary orientation intended by the operator) and 0NQ(the orientation of the path geometry) is less than the threshold a. In this embodiment, there is no directionality associated with the striping directions, e.g., a heading of 90° is equivalent to a heading of 270° for the purposes of striping directions. One way of handling this is to define infill striping angles 0Nnas being within 0° to 180°. If the primary orientation 0striPe > 180°, = Ostripe - 180 °, assuming Ostripe is in the range 0° to 360°. In other embodiments, a direction and orientation may be generated for each infill, such that a heading of 90° is different from to a heading of 270° for the purposes of striping directions, for example. In such a case, the different vehicle work paths for the two heading directions may have a same path geometry but different traversal directions

[0061] The number of allowed striping directions can be predetermined based on knowledge of 0Noand A. For example, for four allowed striping directions (e.g., N / S, NE / SW, EAV, and SE / NW), a = 9074 = 22.5°. The following set can be calculated that includes the angle of each of the A fills, No to NA-I: { 0NO, 0Nt=0NO+ 2a, 0N2=9N0+ 4a, ... , 0NA_J=0NO+2(A-l)a }. This example assumes evenly distributed infill angles, although other arrangements are possible. For purposes of this disclosure, the accessing oftwo or more path definitions infers that two or more allowable striping directions are known for the system such that a path definition can be used corresponding to one of those directions, even if the geometry of the path definition has not yet been generated.

[0062] If the initial infill striping angle 0NQis within ±a of Ostripe, then block 1203 returns ‘yes’ and the initial infill is loaded at block 1204. This allows the user to select autonomous operation and start work as indicated by blocks 1205 and 1206. If block 1203 returns ‘no,’ then an index is incremented at block 1207, and a loop comprising blocks 1208 and 1209 is executed to determine a striping infill angle that is closest aligned to Ostripe, e.g., within ±a of Ostripe. Note that as long as A is known, the number of striping infill angles is known as described above, even if an infill for that direction has not been generated. This case is indicated at blocks 1209 and 1210, in which an infill is created if it was not previously generated. The selected infill Nncan then be loaded 1211 and used at blocks 1205 and 1206.

[0063] The example in FIG. 12 can be extended to incorporate features from other examples. For instance, there may be two or more path definitions for each infill angle 0Nn, such as one infill being offset by a half vehicle width by another infill at the same angle. The start location Lstripe and striping angle Ostripe could both be considered when selecting a path definition. For example, a path definition with a path that passes closest to Lstripe and is most closely aligned with Ostripe could both be selected.

[0064] In other cases, a mowing pattern may use multiple different paths to cover the same work region, e.g., two passes 90° offset from each other for a hatched mowing pattern. In such a case, an initial fill may be selected for a first pass as described above corresponding to Ostripe, then a second infill could be automatically selected for the second pass at Ostripe + A, in which the second infill is offset from the first infill by the offset angle A. In other cases, the first and second infills may be used on different days. In that case, the operator may not remember which direction was worked last. The algorithm could be adapted to select either a path closest aligned to the operator-specified striping angle or to the path at the A offset angle, e.g., 90° offset from the operator-specified striping angle. In this way, the operator need not know or remember the last angle used, but defining the start orientation as one or the other will cause the correct one to be worked.

[0065] In FIG. 13, a block diagram shows components of various apparatuses and systems as described herein. A work vehicle 1300 includes a system controller 1301 adapted to monitor and control various system functions. The system controller 1301 may include one or more processors that are coupled to one or more input / output (I / O) busses1302. The controller 1301 receive various inputs and executes one or more computer programs or applications stored in memory 1304. The memory 1304 may include volatile memory (e.g., dynamic random-access memory) or non-volatile memory (e.g., flash memory). Other hardware may also be coupled to the I / O busses 1302 including sensors1303, power management circuits 1305, batteries 1306, drive controller circuits 1307, drive motor / actuators 1308, work unit controller 1309, work unit motor / actuators 1310, and external data interface 1311 that allows exchanging data with an external computing arrangement, e.g., operations facility 1320.

[0066] The memory 1304 includes computer-readable instructions or applications that, when executed, e.g., by one or more processors, cause the controller 1301 to perform various computations and / or issue commands. That is to say, the controller 1301 and memory 1304 may together define a computing apparatus operable to process input data and generate the desired output to one or more components / devices. For example, the sensors 1303 may include a Global Position System (GPS) receiver and / or wheel encoders that generate signals processed by the controller 1301 (e.g., via autonomous navigation software 1312) to generate speed and steering angle commands to the one or more motors 1308 to cause the drive wheels to rotate (at the same or different speeds and in the same or different directions). In other words, the controller 1301 may control the steering angle and speed of the vehicle 1300.

[0067] The sensors 1303 may include navigation sensors such as cameras, as proximity sensors, ground contact sensors, cameras, LIDAR, Global Navigation Satellite Systems sensors (GNSS), real-time kinetic (RTK) sensors, wheel encoders, radar, LIDAR, contact sensors, beacon detectors, boundary wire detectors, etc. In various embodiments described herein, the vehicle 1300 operates with one or more work implements that perform work tasks, such as cutting turf, aerating turf, spraying turf or other plants, ground leveling, irrigation, fertilization, etc. The work implement may include independent motorsand / or actuators 1310 that drive the implement. A work unit controller 1309 can enable and disable the motor s / actuators 1310, and control aspects such as speed, cut height, etc.[00681 The vehicle 1300 may have the ability to communicate with an operations facility 1320, which may include a data center and / or computing device with similar functionality, e.g., an on-site mobile device. Example functions provided by the operations center include map management 1321, which involves the learning, definition, storage, and retrieval of maps of a work region and work paths within the work region. The work paths may be generated and stored before work commences, e g., after an initial training on work region boundaries and obstacles. In some cases, work paths may be dynamically generated for one or more work sessions. The work paths are used by an autonomous navigation unit 1312. A manual control unit 1313 may also be used to manually control machine operations, as well as provide inputs to communicate information that can imply operator intent.

[0069] The operations facility 1320 may assist in work planning 1322, such as identification of maps and paths for use at a current location, estimating time to completion, monitoring work as it goes on, etc. The operations facility 1320 may have facilities for fleet management 1323, which generally involve monitoring a fleet of vehicles for purposes such as coordination to match a collection of vehicles to a particular job, identification of spares, tracking equipment maintenance, etc.

[0070] In view of the above, it will be readily apparent that the functionality of the controllers of the system may be implemented in any manner known to one skilled in the art. For instance, the memory may include any volatile, non-volatile, magnetic, optical, and / or electrical media, such as a random-access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically-erasable programmable ROM (EEPROM), flash memory, and / or any other digital media.

[0071] The processors used in the controllers may include any one or more of a microprocessor, a controller, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), and / or equivalent discrete or integrated logic circuitry. In some embodiments, the processor may include multiple components, such as any combination of one or more microprocessors, one ormore controllers, one or more DSPs, one or more ASICs, and / or one or more FPGAs, as well as other discrete or integrated logic circuitry. The functions attributed to the controller and / or processor herein may be embodied as software, firmware, hardware, or any combination of these. Certain functionality of the controller may also be performed in the cloud or other distributed computing systems operably connected to the processor.

[0072] The inter-device and intra-device communications may use any combination of wired and wireless communications. Examples of wireless data interfaces include WiFi, Bluetooth, cellular modem, inductive data interface, and NFC. Examples of wired interfaces include Universal Serial Bus (USB), Ethernet, Controller Area Network (CAN), Inter-Integrated Circuit (I2C), and serial line (e.g., RS-232, IEEE 1394).

[0073] In FIG. 14, a flowchart illustrates a method according to an example embodiment. The method involves accessing 1400 a path definition that defines a vehicle work path through a work region. During a staging of an autonomous machine at an arbitrary location within the work region by an operator, a starting pose of the autonomous machine is determined 1401, e.g., using a sensor on the machine. A target navigation start region is determined 1402 (e g., computed using an algorithm) within the work region based on the starting pose. The target navigation start region representing an allowable region within which to start work. One or more start points on the work path within the target navigation start region are selected 1403. The selected start point provides an estimate of an intent of the operator during the staging. The selected start point is used 1404 for performing autonomous work with the autonomous machine along on the vehicle work path.

[0074] In FIG. 15, a flowchart illustrates a method according to another example embodiment. The method involves accessing 1500 (e.g., generating and / or retrieving from data storage) two or more path definitions that define different vehicle work paths through a work region. During a staging of an autonomous machine at an arbitrary location within the work region by an operator, a starting pose of the autonomous machine is determined 1501 (e.g., measured via sensor of the machine). One of the two or more path definitions is selected 1502 based on the starting pose. The selected path definition defines a selected work path that estimates an intent of the operator during the staging. One or more startpoints on the selected work path are selected 1503, and the selected start point is used 1504 for performing autonomous work along on the selected work path.[00751 While the present disclosure is not so limited, an appreciation of various aspects of the disclosure will be gained through a discussion of the specific illustrative aspects provided below. Various modifications of the illustrative aspects, as well as additional aspects of the disclosure, will become apparent herein.

[0076] Example l is a method comprising: during a staging of an autonomous machine in or near a work region by an operator, determining a primary orientation for autonomous work; accessing two or more path definitions that define different vehicle work paths through the work region; selecting one of the two or more path definitions based on the primary orientation; and using the selected path definition to perform the autonomous work.

[0077] Example A2 includes the method of example Al, wherein selecting the path definition based on the primary orientation comprises determining that an angle between the primary orientation and an orientation of a primary part of the selected path definition is less than a threshold. Example A3 includes the method of any previous A example, wherein the different vehicle work paths are alternate paths that cover a same area of the work region. Example A4 includes the method of example A3, wherein at least two of the different vehicle work paths have a same path geometry but different traversal directions. Example A5 includes the method of example A3 or A4, wherein at least two of the different vehicle work paths have a different path geometry, wherein subsequent applications of the different work paths results in contrasting or overlapping pattern in the work region. Example A6 includes the method of example A5, wherein the selected path definition is selected based on being different than a last path definition that was used to autonomously work the work region.

[0078] Example A7 includes the method of any previous A example, further comprising: training the autonomous machine to learn one or more boundaries of the work region; auto-generating a first one of the two or more path definitions to conform to the one or more boundaries; and storing first path data defining the first path definition for later retrieval and use by the autonomous machine. Example A8 includes the method ofexample A7, wherein the one or more boundaries comprise both containment boundaries and exclusion boundaries.[00791 Example A9 includes the method of example any previous A example, further comprising determining a start heading of the autonomous machine during the staging, the start heading being used as the primary orientation. Example A10 includes the method of example A9, further comprising determining a start location of the autonomous machine during the staging. Example Al 1 includes the method of example A10, wherein the two or more path definitions comprise path segments proximate the start location that are oriented in different path directions, the selected path being selected based on the path segment of the selected path being closest-aligned with the start heading.

[0080] Example A12 includes the method of example A10 or Al l, wherein the two or more path definitions comprise path segments proximate the start location that are oriented in different path directions and wherein none of the path segments are aligned to the start heading within a threshold, the method further comprising auto-generating an additional path definition having at least one path segment aligned with the start heading proximate the start location, the additional path definition used to autonomously work the work region. Example Al 3 includes the method of example A 12, wherein an angle of the path segment is selected from a predefined set of angles, the angle being within the threshold from the start heading.

[0081] Example A14 includes the method of any previous A example, further comprising predefining a set of angles, the two or more path definitions having two or more different primary angles set at different angles of the set of angles. Example Al 5 includes the method of any previous A example, wherein determining the primary orientation comprises moving the autonomous machine along a manually defined path, the manually defined path defining the primary orientation. Example Al 6 includes the method of any previous A example, wherein determining the primary orientation comprises defining the primary orientation on a mobile device map. Example Al 7 includes the method of any previous A example, further comprising: determining a boundary of the work region based on a first operator input; determining an initial angle based on a second operator input; auto-generating a first one of the two or more path definitions to conform tothe boundary and the initial angle; and storing first path data defining the first path definition for later retrieval and use by the autonomous machine.[00821 Example B 17 is a method comprising: accessing two or more path definitions that define different vehicle work paths through a work region; during a staging of an autonomous machine at an arbitrary location within the work region by an operator, determining a starting pose of the autonomous machine; selecting one of the two or more path definitions based on the starting pose, the selected path definition defining a selected work path that estimates an intent of the operator during the staging; and using the selected work path for performing autonomous work.

[0083] Example B 18 includes the method of example B 17, wherein the different vehicle work paths are represented as geometric loops. Example B19 includes the method of example Bl 8, further comprising traversing the geometric loop of the selected work path using the starting pose as a start point and an end point.

[0084] Example B20 includes the method of example B 18 or B 19, wherein the one or more path definitions each comprise: an infill path that fills an area of the work region; a clean-up path that follows a boundary of the work region; and a transition path connecting the infill path and the clean-up path. Example B21 includes the method of example B20, wherein the infill path of each of the one or more path definitions comprise parallel lines, the selected path definition being determined based on an the angle between the parallel lines and a heading of the starting pose of the autonomous machine.

[0085] Example B22 includes the method of any previous B example, wherein the one or more path definitions comprise parallel lines, the selected path definition being determined based on an angle between the parallel lines and a heading of the starting pose of the autonomous machine. Example B23 includes the method of any previous B example, wherein if none of the one or more path definitions estimates the intent of the operator, the method further comprises generating a new path definition that defines a new vehicle work path through the work region that estimates the intent of the operator.

[0086] Example B24 is a system, comprising: the autonomous machine set forth in any previous B example, the autonomous machine comprising a controller operable to use the selected path definition for performing the autonomous work; and an operations facilitycoupled to exchange data with the autonomous machine, the operations facility comprising a processor operable to select the selected path definition based on the starting pose.Example B25 includes the system of example B24, wherein the processor of the operations facility is further operable to generate the two or more path definition and communicate at least the selected path definition to the autonomous machine.

[0087] Example C26 is method comprising: during a staging of an autonomous machine in or near a work region by an operator, determining a primary orientation for autonomous work; accessing a predefined path definition that defines a vehicle work path through the work region, a geometry of the path definition defining a predefined primary direction; determining that the predefined path definition does not satisfy a condition, the condition comprising one or more of: the predefined primary direction being aligned with the primary orientation within a threshold; and a work requirement based on the predefined primary direction; auto-generating an additional path definition with an additional primary direction that satisfies the condition; and using the additional path definition to perform the autonomous work.

[0088] Example C27 includes the method of example C26, further comprising storing the additional path definition with the predefined path definition for later retrieval and use by the autonomous machine. Example C28 includes the method of example C26 or C27, wherein the work requirement comprises a requirement that primary path directions between subsequent work sessions vary. Example C29 includes the method of example C26, C27, or C28, further comprising predefining a set of angles, the predefined primary direction and the additional primary direction being set at different angles of the set of angles.

[0089] It is noted that the terms “have,” “include,” “comprises,” and variations thereof, do not have a limiting meaning, and are used in their open-ended sense to generally mean “including, but not limited to,” where the terms appear in the accompanying description and claims. Further, “a,” “an,” “the,” “at least one,” and “one or more” are used interchangeably herein. Moreover, relative terms such as ’’left,” “right,” “front,” “fore,” “forward,” “rear,” “aft,” “rearward,” “top,” “bottom,” “side,” “upper,” “lower,” “above,” “below,” “horizontal,” “vertical,” and the like may be used herein and, ifso, are from the perspective shown in the particular figure, or while the machine is in an operating configuration. These terms are used only to simplify the description, however, and not to limit the interpretation of any embodiment described. As used herein, the terms “determine” and “estimate" may be used interchangeably depending on the particular context of their use, for example, to determine or estimate a position or pose of a vehicle, boundary, obstacle, etc.

[0090] Further, it is understood that the description of any particular element as being connected to or coupled to another element can be directly connected or coupled, or indirectly coupled / connected via intervening elements.

[0091] Unless otherwise indicated, all numbers expressing feature sizes, amounts, and physical properties used in the specification and claims are to be understood as being modified in all instances by the term “about.” Accordingly, unless indicated to the contrary, the numerical parameters set forth in the foregoing specification and attached claims are approximations that can vary depending upon the desired properties sought to be obtained by those skilled in the art utilizing the teachings disclosed herein. The use of numerical ranges by endpoints includes all numbers within that range (e.g., 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.80, 4, and 5) and any range within that range.

[0092] The various embodiments described above may be implemented using circuitry, firmware, and / or software modules that interact to provide particular results. One of skill in the arts can readily implement such described functionality, either at a modular level or as a whole, using knowledge generally known in the art. For example, the flowcharts and control diagrams illustrated herein may be used to create computer-readable instructions / code for execution by a processor. Such instructions may be stored on a non- transitory computer-readable medium and transferred to the processor for execution as is known in the art. The structures and procedures shown above are only a representative example of embodiments that can be used to provide the functions described hereinabove.

[0093] Note that any components described herein using terms such as “processor,” “controller,” “logic circuit,” “CPU,” or the like may be implemented using a plurality of discrete units operating together. For example, a processer that performs a series of steps or operations may be construed as two or more processors operating cooperatively to performthe steps. Similarly, other processing hardware such as memory and input-output may perform the described functions with multiple discrete units operating cooperatively or being coordinated by another unit, e.g., by a central processor or processors.

[0094] The foregoing description of the example embodiments has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the embodiments to the precise form disclosed. Many modifications and variations are possible in light of the above teaching. Any or all features of the disclosed embodiments can be applied individually or in any combination and are not meant to be limiting, but purely illustrative. It is intended that the scope of the invention be limited not with this detailed description, but rather determined by the claims appended hereto.

Claims

CLAIMS:

1. A method comprising: during a staging of an autonomous machine in or near a work region by an operator, determining a primary orientation for autonomous work; accessing two or more path definitions that define different vehicle work paths through the work region; selecting one of the two or more path definitions based on the primary orientation; and using the selected path definition to perform the autonomous work.

2. The method of claim 1, wherein selecting the path definition based on the primary orientation comprises determining that an angle between the primary orientation and an orientation of a primary part of the selected path definition is less than a threshold.

3. The method of claim 1 or 2, wherein the different vehicle work paths are alternate paths that cover a same area of the work region.

4. The method of claim 3, wherein at least two of the different vehicle work paths have a same path geometry but different traversal directions.

5. The method of claim 3, wherein at least two of the different vehicle work paths have a different path geometry, wherein subsequent applications of the different work paths results in contrasting or overlapping pattern in the work region.

6. The method of claim 5, wherein the selected path definition is selected based on being different than a last path definition that was used to autonomously work the work region.

7. The method of any previous claim, further comprising:training the autonomous machine to learn one or more boundaries of the work region; auto-generating a first one of the two or more path definitions to conform to the one or more boundaries; and storing first path data defining the first path definition for later retrieval and use by the autonomous machine.

8. The method of claim 7, wherein the one or more boundaries comprise both containment boundaries and exclusion boundaries.

9. The method of any previous claim, further comprising determining a start heading of the autonomous machine during the staging, the start heading being used as the primary orientation.

10. The method of claim 9, further comprising determining a start location of the autonomous machine during the staging.

11. The method of claim 10, wherein the two or more path definitions comprise path segments proximate the start location that are oriented in different path directions, the selected path being selected based on the path segment of the selected path being closest- aligned with the start heading.

12. The method of claim 10, wherein the two or more path definitions comprise path segments proximate the start location that are oriented in different path directions and wherein none of the path segments are aligned to the start heading within a threshold, the method further comprising auto-generating an additional path definition having at least one path segment aligned with the start heading proximate the start location, the additional path definition used to autonomously work the work region.

13. The method of claim 12, wherein an angle of the path segment is selected from a predefined set of angles, the angle being within the threshold from the start heading.

14. The method of any previous claim, further comprising predefining a set of angles, the two or more path definitions having respective two or more different primary angles set at different angles of the set of angles, wherein the path segment is selected based on minimizing a difference between a heading of the primary orientation and the primary angle of the selected path definition.

15. The method of any previous claim, wherein determining the primary orientation comprises moving the autonomous machine along a manually defined path, the manually defined path defining the primary orientation.

16. The method of any one of claims 1-14, wherein determining the primary orientation comprises defining the primary orientation on a mobile device map.

17. The method of any previous claim, wherein selected path definition estimates an intent of the operator during the staging.

18. The method of any previous claim, further comprising: determining a boundary of the work region based on a first operator input; determining an initial angle based on a second operator input; auto-generating a first one of the two or more path definitions to conform to the boundary and the initial angle; and storing first path data defining the first path definition for later retrieval and use by the autonomous machine.

Images