Grade Control Cleanup Pass Using Cost Optimization

Abstract

A computer-implemented method for determining a cleanup pass profile for a machine implement is provided. The method may include identifying a pass target extending from a first end to a second end along a work surface, determining a plurality of primitives of a cleanup pass profile extending between the first end and the second end where each primitive may be defined based at least partially on volume constraints, determining a cost value associated with moving the machine implement along the cleanup pass profile based on an optimization cost function, and adjusting the volume constraints associated with one or more of the primitives to minimize the cost value.

Description

TECHNICAL FIELD

The present disclosure relates generally to controlling machines, and more particularly, to systems and methods for determining cleanup pass profiles for semi-autonomous and autonomous machines using cost optimization.

BACKGROUND

Machines such as, for example, track-type tractors, dozers, motor graders, wheel loaders, and the like, are used to perform a variety of tasks. For example, these machines may be used to move material and/or alter work surfaces at a worksite. The machines may be manned machines, but may also be autonomous or semi-autonomous vehicles that perform these tasks in response to commands remotely or locally generated as part of a work plan for the machines. The machines may receive instructions in accordance with the work plan to perform operations, including digging, loosening, carrying, and any other manipulation of materials at the worksite.

It may be desirable to ensure that the machines perform these operations such that the materials are moved in an efficient manner. More particularly, in repetitive operations, it may be especially desirable to ensure that the locations at which the machines begin to alter the work surface, or the profiles along which the machines alter the work surface, are selected in a way that maximizes efficiency and productivity. Some conventional systems, such as disclosed in U.S. Pat. Appl. Publ. No. 2014/0012404, published on Jan. 9, 2014 and entitled “Methods and Systems for Machine Cut Planning,” plan cut locations based on predetermined cut volume estimations. While such techniques may greatly assist in the planning processes and the overall excavation, there is still room for improvement.

A standard cut profile in autonomous dozing is generally composed of three regions, including a blade-in-air region, a blade-load region, and a blade-carry region. In the blade-in-air region, a dozer is typically reversing after a cut and positioning a blade implement to make contact with the work surface. Once contact is made with the work surface and a cut is initiated, the blade is loaded with material in the blade-load region and generally moved downward to a target carry surface. In the blade-carry region, the blade carries the loaded material to a crest of the worksite. As this process is repeated, the work surface elevation gradually changes and the profile of the blade-load region is updated accordingly. However, autonomous carry passes often adjust the blade height while in the blade-carry region which can result in unwanted deviations from the planned profile.

Theoretically, conventional cut and carry passes, along with occasional ripping passes, may be repeated to execute clean passes according to the planned profile and avoid unwanted deviations. In actual practice, however, cut and carry passes may deviate from the planned profile due to factors such as hard soil, insufficient ripping, degradations in position estimation, hump building, large rocks, boulders or other embedded obstacles, and the like. Limitations in the actual process of planning for conventional cut and carry passes are also factors. For instance, conventional processes are limited to profiles formed using S-shaped Gaussian curves, which cannot sufficiently adapt to negative volumes or valleys in the terrain that dip below the target profile, bumps in the terrain that extend above the pass target, or the like.

Accordingly, there is a need for grade control or cleanup passes that can reduce inconsistencies in the terrain, minimize operator involvement, and help improve productivity of the overall excavation. Furthermore, there is a need for cleanup pass profiling systems and methods that provide more versatile means for correcting surface irregularities, such as by shaving, snaking or otherwise cutting bumps and/or small valleys, but at minimal cost. The present disclosure is directed at addressing one or more of the inefficiencies and disadvantages set forth above. However, it should be appreciated that the solution of any particular problem is not a limitation on the scope of this disclosure or of the attached claims except to the extent express noted.

SUMMARY OF THE DISCLOSURE

In one aspect of the present disclosure, a computer-implemented method for determining a cleanup pass profile for a machine implement is provided. The method may include identifying a pass target extending from a first end to a second end along a work surface, determining a plurality of primitives of a cleanup pass profile extending between the first end and the second end where each primitive may be defined based at least partially on volume constraints, determining a cost value associated with moving the machine implement along the cleanup pass profile based on an optimization cost function, and adjusting the volume constraints associated with one or more of the primitives to minimize the cost value.

In another aspect of the present disclosure, a control system for determining a cleanup pass profile for a machine implement is provided. The control system may include a memory configured to retrievably store one or more algorithms, and a controller in communication with the memory. The controller, based on the one or more algorithms, may be configured to at least identify a pass target extending from a first end to a second end along a work surface, determine a plurality of primitives of a cleanup pass profile extending between the first end and the second end, determine a cost value associated with moving the machine implement along the cleanup pass profile based on an optimization cost function, and adjust volume constraints associated with one or more of the primitives to minimize the cost value.

In yet another aspect of the present disclosure, a controller for determining a cleanup pass profile for a machine implement is provided. The controller may include a pass target identification module configured to identify a pass target extending from a first end to a second end along a work surface, a cleanup pass profile module configured to determine a plurality of primitives of a cleanup pass profile extending between the first end and the second end, and a cost optimization module configured to determine a cost value associated with moving the machine implement along the cleanup pass profile based on an optimization cost function, and adjust volume constraints associated with one or more of the primitives to minimize the cost value.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a pictorial illustration of an exemplary disclosed worksite;

FIG. 2 is a diagrammatic illustration of an exemplary control system that may be used at a worksite;

FIG. 3 is a diagrammatic illustration of an exemplary controller that may be used at a worksite;

FIG. 4 is a diagrammatic illustration of an exemplary cleanup pass profile that may be generated by a control system of the present disclosure using cost optimization; and

FIG. 5 is a flowchart depicting an exemplary disclosed method that may be performed by a control system of the present disclosure.

DETAILED DESCRIPTION

Although the following sets forth a detailed description of numerous different embodiments, it should be understood that the legal scope of protection is defined by the words of the claims set forth at the end of this patent. The detailed description is to be construed as exemplary only and does not describe every possible embodiment since describing every possible embodiment would be impractical, if not impossible. Numerous alternative embodiments could be implemented, using either current technology or technology developed after the filing date of this patent, which would still fall within the scope of the claims defining the scope of protection.

It should also be understood that, unless a term is expressly defined herein, there is no intent to limit the meaning of that term, either expressly or by implication, beyond its plain or ordinary meaning, and such term should not be interpreted to be limited in scope based on any statement made in any section of this patent (other than the language of the claims). To the extent that any term recited in the claims at the end of this patent is referred to herein in a manner consistent with a single meaning, that is done for sake of clarity only so as to not confuse the reader, and it is not intended that such claim term be limited, by implication or otherwise, to that single meaning.

Referring now to FIG. 1, one exemplary worksite 100 is illustrated with one or more machines 102 performing predetermined tasks. The worksite 100 may include, for example, a mine site, a landfill, a quarry, a construction site, or any other type of worksite. The predetermined task may be associated with altering the geography at the worksite 100, such as a dozing operation, a grading operation, a leveling operation, a bulk material removal operation, or any other type of operation that results in geographical modifications within the worksite 100. The machines 102 may be mobile machines configured to perform operations associated with industries related to mining, construction, farming, or any other industry known in the art. The machines 102 depicted in FIG. 1, for example, may embody earth moving machines, such as dozers having blades or other work tools or implements 104 movable by way of one or more actuators 106. The machines 102 may also include manned machines or any type of autonomous or semi-autonomous machines.

The overall operations of the machines 102 and the machine implements 104 within the worksite 100 may be managed by a control system 108 that is at least partially in communication with the machines 102. Moreover, each of the machines 102 may include any one or more of a variety of feedback devices 110 capable of signaling, tracking, monitoring, or otherwise communicating relevant machine information to the control system 108. For example, each machine 102 may include a locating device 112 configured to communicate with one or more satellites 114, which in turn, may communicate to the control system 108 various information pertaining to the position and/or orientation of the machines 102 relative to the worksite 100. Each machine 102 may additionally include one or more implement sensors 116 configured to track and communicate position and/or orientation information of the implements 104 to the control system 108.

The control system 108 may be implemented in any number of different arrangements. For example, the control system 108 may be at least partially implemented at a command center 118 situated locally or remotely relative to the worksite 100 with sufficient means for communicating with the machines 102, for example, via satellites 114, or the like. Additionally or alternatively, the control system 108 may be implemented using one or more computing devices 120 with means for communicating with one or more of the machines 102 or one or more command centers 118 that may be locally and/or remotely situated relative to the worksite 100. In still further alternatives, the control system 108 may be implemented on-board any one or more of the machines 102 that are also provided within the worksite 100. Other suitable modes of implementing the control system 108 are possible and will be understood by those of ordinary skill in the art.

Using any of the foregoing arrangements, the control system 108 may generally be configured to monitor the positions of the machines 102 and/or machine implements 104 relative to the worksite 100 and a predetermined target operation, and provide instructions for controlling the machines 102 and/or machine implements 104 in an efficient manner in executing the target operation. In certain embodiments, the machines 102 may be configured to excavate areas of a worksite 100 according to one or more predefined excavation plans. The excavation plans can include, among other things, determining the location, size, and shape of a plurality of cuts into an intended work surface 122 at the worksite 100 along one or more slots 124. In such embodiments, the control system 108 may be used to plan not only the overall excavation, but also to plan intermittent grade control or cleanup passes within the slots 124 or any other areas of the work surface 122. For a given work surface 122 and pass target, for instance, the control system 108 may generate a cleanup pass profile best suited to remove surface irregularities, such as smaller bumps and valleys in the work surface 122, which may adversely affect the autonomous or semi-autonomous performance of the overall excavation. Although described in connection with grade control or cleanup pass planning and profiling, the control system 108 may similarly be employed in conjunction with other types of tasks.

Turning to FIG. 2, one exemplary embodiment of a control system 108 that may be used in conjunction with the worksite 100 and the machines 102 of FIG. 1 is diagrammatically provided. As shown, the control system 108 may generally include, among other things, a controller 126, a memory 128, and a communications device 130. More specifically, the controller 126 may be configured to operate according to one or more algorithms that are retrievably stored within the memory 128. The memory 128 may be provided on-board relative to the controller 126, external to the controller 126, or otherwise in communication therewith. The communications device 130 may be configured to enable the controller 126 to communicate with one or more of the machines 102, and receive information pertaining to the position and/or orientation of the machines 102 and the machine implements 104, for example, via satellites 114, or any other suitable means of communication. Moreover, the controller 126 may be implemented using any one or more of a processor, a microprocessor, a microcontroller, or any other suitable means for executing instructions stored within the memory 128. Additionally, the memory 128 may include non-transitory computer-readable medium or memory, such as a disc drive, flash drive, optical memory, read-only memory (ROM), or the like.

As further shown in FIG. 3, the controller 126 may be configured to at least determine a grade control or cleanup pass profile at a worksite 100 according to one or more preprogrammed algorithms which may be generally categorized into, for example, a work surface identification module 132, a pass target identification module 134, a cleanup pass profile module 136, and a cost optimization module 138. With further reference to the exemplary diagram of FIG. 4, the work surface identification module 132 may configure the controller 126 to initially identify the work surface 122 to be worked on, such as in terms of position relative to the worksite 100, position relative to the machines 102, elevation, slope, volume of material moved, removed or remaining, terrain composition, or any other relevant geographical profile. As shown for example in FIG. 4, a given work surface 122 may generally be defined as the section of terrain along a slot 124 extending between an alignment gap 140 at a first end and a crest 142 at the second end. Information pertaining to the work surface 122 and/or changes thereto may be communicated to the controller 126 via manual entries, preprogrammed entries, periodically updated entries, real-time entries, or any combination thereof. Moreover, the work surface identification module 132 may configure the controller 126 to map the work surface 122 in two-dimensional formats, such as shown in FIG. 4, or in other alternatives, in three-dimensional formats.

The pass target identification module 134 of FIG. 3 may configure the controller 126 to identify the carry surface or pass target 144 that is ultimately desired. As with the work surface identification module 132, the pass target identification module 134 may identify the pass target in terms of location or position relative to the worksite 100, position relative to the machines 102, position relative to the work surface 122, elevation, slope, volume differential with the work surface 122, terrain composition, or any other relevant geographical profile. Additionally, the pass target 144 may generally extend the length of the work surface 122 between the alignment gap 140 and the crest 142. The pass target 144 may be identified using any number of different techniques. As shown for instance in FIG. 4, the pass target 144 may be identified or defined based on a two-dimensional user-defined curve that is positioned, superimposed or otherwise mapped relative to the work surface 122. Moreover, information defining the pass target 144 may be manually input, programmed or preprogrammed into the controller 126. In other alternative embodiments, the pass target 144 may be identified based on a two-dimensional cross-section or slice of a three-dimensional model of the pass target 144. In still further modifications, the controller 126 may be configured to identify the work surface 122 and the pass target 144 using three-dimensional models, or the like.

With both the work surface 122 and the pass target 144 identified, the cleanup pass profile module 136 of FIG. 3 may configure the controller 126 to determine a plurality of geometric primitives 146 that can ultimately be combined or adjoined at its endpoints 148 to begin generating a cost-optimized cleanup pass profile 150 for the given work surface 122 and pass target 144. Each primitive 146 may include a line, a parabola, a cubic, a polynomial, a Gaussian curve, an exponential function, or any geometric primitive that can be provided along the work surface 122 generally extending from the alignment gap 140 toward the crest 142 as shown in FIG. 4. Furthermore, each primitive 146 may be defined or constrained based on the slope and/or elevation of other adjoining primitives 146, adjoining work surfaces 122 and pass targets 144. The elevation or curve function of a given primitive 146 may be provided as, for example,

y(x)=c₀+c₁x+ . . . +c_mkx^mk  (1)

and the slope or derivative of that function may be provided as, for example,

$\begin{array}{cc}\frac{\partial y}{\partial x}=c_1+2c_2x+\dots +m_kc_{m_k}x^{m_k-1}& \left(2\right)\end{array}$

where m denotes the order of the polynomial making up the kth primitive, and c denotes the unknown polynomial coefficients to be resolved.

The cleanup pass profile module 136 may also define each primitive 146 based on volume constraints which may be derived using, for example, volume differentials between the work surface 122 and the pass target 144, and load limits of the associated machines 102 and machine implements 104. Volume differentials may indicate the volume of material that needs to be moved for a given pass, while the machine or implement load limits may suggest the depth, elevation, slope or other parameters with which the machine 102 and the implement 104 should operate for the given pass. In one embodiment, the polynomial function of each primitive 146 may be defined by the volume constraint

$\begin{array}{cc}{w}_b·{\int }_{d_{k-1}}^{d_k}y\left(x\right)x=w_b·\left(c_0x+\frac{1}{2}c_1x^2+\dots +\frac{1}{m_k+1}c_{m_k}x^{m_k+1}|\begin{array}{c}{d}_k\\ d_{k-1}\end{array}\right)& \left(3\right)\end{array}$

where w_b may be a parameter or dimension of the implement 104 used, such as blade width, or the like, and where d_k indicates the end of the kth primitive 146 while d_k−1 indicates the start of the kth primitive 146. Using mathematical relationships between a sufficient set of functions or constraints, such as those of equations (1)-(3), the polynomial coefficients that define each primitive 146 may be determined. For instance, in a pass with n number of primitives 146, solving a set of 2(n−1) equations may provide coefficients that will match the slope and elevation at the endpoints 148 of adjoining primitives 146, and solving a set of four equations may provide coefficients that will match the slope and elevation at the relevant endpoints 148 of the first and final primitives 146 to the work surface 122 or the pass target 144. Additionally, solving a corresponding set of (n−2) equations may provide coefficients that will constrain the shape of each primitive 146 to the target volume to be moved.

As shown for example in the embodiment of FIG. 4, the cleanup pass profile module 136 may form each primitive 146 using a curve defined by a polynomial function that is constrained in terms of both slope and elevation, as well as in accordance with volume constraints. In particular, the slope and elevation of the endpoints 148 of the first primitive 146-1 may be configured to match those of the work surface 122 at the alignment gap 140 adjacent thereto as well as those of the second primitive 148-2. Correspondingly, the slope and elevation of the endpoints 148 of the final primitive 146-6 may be configured to match those of the crest 142 adjacent thereto as well as those of the fifth primitive 146-5. The slope and elevation of the adjoined endpoints 148 between the intermediate primitives 146-2, 146-3, 146-4, 146-5 may similarly be configured to match. As shown, each of the primitives 146 may be further defined based on volume constraints which help shape each primitive 146 and the overall cleanup pass profile 150. For instance, the first primitive 146-1 and the second primitive 146-2 together may be designated as a cut region with a planned cut volume of 100% of the machine load. The third primitive 146-3 may be configured as a fill region with a planned fill volume of 80% of the load capacity. The fourth primitive 146-4 and the fifth primitive 146-5 may be configured together as cut regions with a 65% planned cut volume. The final primitive 146-6 may be designated as a cut region with a 15% planned cut volume.

Additionally, the cost optimization module 138 of FIG. 3 may further configure the controller 126 to optimize the cleanup pass profile 150 according to one or more optimization cost functions preprogrammed therein. In general, the optimization cost function may assess a particular cleanup pass profile 150 and the primitives 146 thereof in terms of the cost associated with operating or moving the machine 102 and the machine implement 104 along the cleanup pass profile 150. For example, the optimization cost function may determine a cost value based on an assessment of one or more parameters associated with the cleanup pass profile 150, including curvature, curvature limit, slope, slope limit, implement load, implement load average, maximum implement load limit, minimum implement load limit, crest volume, cut depth relative to pass target, cut depth relative to work surface, volume differential between the pass target 144 and the work surface 122, and the like. In one embodiment, the optimization cost function may increment the cost value for each parameter of the cleanup pass profile 150 that exceeds a corresponding desired or target threshold. Furthermore, one or more of the parameters may be weighted according to its relevance to the overall cost of performing the cleanup pass such that, for example, the cost value is incremented in proportion to the relevance of the parameter that exceeds a corresponding threshold.

In addition to determining the cost value associated with each cleanup pass profile 150, the cost optimization module 138 may further configure the controller 126 to adjust one or more of the primitives 146 of the cleanup pass profile 150 to determine the most cost effective cleanup pass profile 150 for the given circumstances. In particular, the controller 126 may be configured to adjust one or more of the volume constraints of the primitives 146 to determine one or more variations of the original cleanup pass profile 150, calculate the cost value for each of those variations, and select the variation having the lowest cost value as the final cleanup pass profile 150 to be performed. The controller 126 may adjust the volume constraint of one or more primitives 146, such as in the form of equation (3), using a gradient descent algorithm, a genetic algorithm, a Nelder-Mead algorithm, or any other technique or algorithm commonly used in the art. While adjusting the primitives 146, the controller 126 may be configured to ensure certain variables remain substantially constant, such as the total volume differential between the work surface 122 and the original cleanup pass profile 150, the elevation and slope of each endpoint 148 in the original cleanup pass profile 150, and the like. In alternative embodiments, the controller 126 may adjust the elevation of the endpoints 148 of the original cleanup pass profile 150, the location of the endpoints 148, the number of endpoints 148, and the like.

After exhausting all feasible variations of the original cleanup pass profile 150, and after determining a cost value for each of the original cleanup pass profile 150 and variations thereof, the cost optimization module 138 may configure the controller 126 to compare the cost values and select the group of primitives 146 exhibiting the lowest cost value as the final cleanup pass profile 150. Specifically, the controller 126 may interrelate the polynomial or other curve functions associated with the individual primitives 146, such as via one or more mathematical relationships therebetween, in a manner which adjoins the endpoints 148 of the primitives 146. The controller 126 may further digitalize or otherwise translate functions pertaining to the resulting cleanup pass profile 150 into the appropriate instructions for execution by one or more of the machines 102 and/or implements 104 within the worksite 100. In particular, the instructions corresponding to the cleanup pass profile 150 may be transmitted by the communications device 130 to the appropriate machines 102 or implements 104, which may in turn, execute the cleanup pass accordingly. Other variations and modifications to the algorithms or methods will be apparent to those of ordinary skill in the art. One exemplary algorithm or method by which the controller 126 may be operated to determine a grade control or cleanup pass profile 150 based on cost optimization is discussed in more detail below.

INDUSTRIAL APPLICABILITY

In general, the present disclosure sets forth methods, devices and systems for planning and performing grade control or cleanup passes where there are motivations to improve productivity and efficiency. Although applicable to any type of machine, the present disclosure may be particularly applicable to autonomously or semi-autonomously controlled dozing machines where the dozing machines are controlled along particular travel routes within a worksite to excavate materials. Moreover, the present disclosure may improve the overall excavation process by enabling more versatile and more cost effective grade control or cleanup passes. Furthermore, by providing optimized cleanup pass profiles that can be autonomously or semi-autonomously executed, unwanted irregularities in a given work surface may be efficiently corrected and deviations typically caused thereby may be significantly reduced.

Turning now to FIG. 5, one exemplary algorithm or computer-implemented method 152 for determining a cleanup pass profile 152 is diagrammatically provided, according to which, for example, the control system 108 and the controller 126 may be configured to operate. As shown, the controller 126 may initially determine whether a grade control or cleanup pass is needed, such as by manual or autonomous means. For instance, a cleanup pass may be manually triggered in response to operator input remotely or locally entered via any one or more of the machines 102, command centers 118, computing devices 120, and the like. Alternatively, a cleanup pass may be autonomously triggered, for example, at predefined intervals of time and/or at predefined checkpoints pertaining to the geographical state of work surface 122. Predefined checkpoints may be defined based on any combination of the length of the given pass, the relative elevations of the alignment gap 140 and the crest 142, the slope, the volume of material moved, removed or remaining, and the like. In further alternatives, the control system 108 and the controller 126 may be configured to autonomously assess whether a cleanup pass is appropriate, for instance, based on any deviations in the tracked progress, position and/or orientation of the work machines 102 and machine implements 104.

If no trigger or request for a cleanup pass is detected, the controller 126 may continue monitoring for such triggers while resuming normal cut operations. If a valid request for a cleanup pass is determined, the controller 126 may begin planning a cleanup pass profile 150 that is most appropriate for the given work surface 122 and pass target 144 according to the algorithm or method 152 shown in FIG. 5. According to block 152-1, for example, the controller 126 may be configured to initially identify the work surface 122, such as in terms of position relative to the worksite 100, position relative to the machines 102, elevation, slope, volume of material moved, removed or remaining, terrain composition, or any other relevant geographical profile thereof For a work surface 122 provided along a slot 124, as shown for instance in FIGS. 1 and 4, the controller 126 may additionally identify the locations of the alignment gap 140 and the crest 142. Moreover, the controller 126 may be configured to receive profile information relating to the work surface 122 and/or changes thereto via manual user input, preprogrammed input, periodically updated input, real-time input, or combinations thereof.

Once information regarding the work surface 122 has been sufficiently identified, mapped or otherwise obtained, the controller 126 may further identify the pass target 144 according to block 152-2 of FIG. 5. Specifically, the controller 126 may be configured to identify the pass target 144 in terms of position relative to the worksite 100, position relative to the machines 102, position relative to the work surface 122, elevation, slope, volume differential with the work surface 122, terrain composition, or any other relevant geographical profile. In general, the pass target 144 may extend the length of the work surface 122 between the alignment gap 140 and the crest 142. While the pass target 144 may be identified using any number of different techniques, the controller 126 may identify or define the pass target 144 based on a two-dimensional curve that is positioned, superimposed or otherwise mapped relative to the work surface 122, as shown for example in FIG. 4. Information regarding the pass target 144 may be manually input, programmed or preprogrammed into the controller 126, or alternatively, identified based on a two-dimensional cross-section or slice of a three-dimensional model of the pass target 144. In other alternatives, the controller 126 may be configured to identify the work surface 122 and the pass target 144 using three-dimensional models.

According to block 152-3 of FIG. 5, the controller 126 may determine a plurality of geometric primitives 146 that can be combined to form a cost-optimized cleanup pass profile 150. For instance, the controller 126 may define the slope and/or elevation of each endpoint 148 of each primitive 146 to match those of adjoining work surfaces 122, pass targets 144, other primitives 146, or the like. The controller 126 may also define each primitive 146 based on volume constraints, which may further be derived based on volume differentials between the work surface 122 and the pass target 144, load limits of the machine 102 and/or implement 104, or the like. Volume differentials may indicate the volume of terrain material that needs to be moved for a given pass, while the machine or implement load limits may suggest the depth, elevation, slope or other characteristics with which the machine 102 and the implement 104 may operate to provide for an efficient cleanup pass. Based on such constraints, the controller 126 may generate relationships, such as between equations (1)-(3), for defining the slope and elevation per primitive 146 or group of primitives 146, as well as the target cut or fill volume per primitive 146 or group of primitives 146. For primitives 146 based on polynomial functions, for example, the controller 126 may be configured to solve for the unknown polynomial coefficients in a given set of equations which will define the specific shape, slope and elevation of each primitive 146, as well as the target volume of material to be cut or filled per primitive 146.

Once an initial cleanup pass profile 150 is determined, the controller 126 in block 152-4 of FIG. 5 may determine the cost value, or the cost associated with operating or moving the machine 102 and/or machine implement 104 along the cleanup pass profile 150. The cost value may be assessed based on an optimization cost function that is preprogrammed in the memory 128 of the control system 108, or in any other memory that is accessible to the controller 126, and calculated based on an assessment of one or more parameters associated with the cleanup pass profile 150. For example, the optimization cost function may assess the cost associated with the cleanup pass profile 150 in terms of curvature, curvature limit, slope, slope limit, implement load, implement load average, maximum implement load limit, minimum implement load limit, crest volume, cut depth relative to pass target, cut depth relative to work surface, volume differential between the pass target 144 and the work surface 122, or the like. The controller 126 may also increment the cost value for each parameter of the cleanup pass profile 150 that exceeds a corresponding desired or target threshold. The controller 126 may additionally apply weights to the parameters according to its relevance to the overall cost of performing the cleanup pass such that, for example, the cost value is incremented in proportion to the relevance of the parameter that exceeds a corresponding threshold.

In addition to determining the cost value associated with each cleanup pass profile 150, the controller 126 in block 152-5 may further adjust one or more of the primitives 146 of the cleanup pass profile 150 to determine the most cost effective cleanup pass profile 150 for the given circumstances. In particular, the controller 126 may adjust one or more of the volume constraints of the primitives 146 to create different variations of the original cleanup pass profile 150 that may produce comparable results in terms of the total volume of terrain or material that is moved within the worksite 100. The controller 126 may adjust the volume constraint of one or more primitives 146, such as in the form of equation (3) above, using a gradient descent algorithm, a genetic algorithm, a Nelder-Mead algorithm, or any other technique or algorithm commonly used in the art. While adjusting the primitives 146, the controller 126 may be configured to ensure that certain variables remain substantially constant, such as the total volume differential between the work surface 122 and the original cleanup pass profile 150, the elevation and slope of each endpoint 148 in the original cleanup pass profile 150, and the like. In alternative embodiments, the controller 126 may adjust the elevation of the endpoints 148 of the original cleanup pass profile 150, the location of the endpoints 148, the number of endpoints 148, and the like.

In block 152-6 of FIG. 5, the controller 126 may be configured to calculate the cost value for each new variation of the original cleanup pass profile 150 proposed in block 152-5. For instance, the controller 126 may assess the cost value for each variation of the cleanup pass profile 150 according to the optimization cost function used in block 152-4. Furthermore, data or functions defining the original cleanup pass profile 150 and each variation thereof, as well as the cost values associated therewith may be at least temporarily stored in memory 128 for additional assessment. According to block 152-7, for example, the controller 126 may be configured to compare the cost value of the original cleanup pass profile 150 to the cost value of each variation thereof to determine the group of primitives 146 exhibiting the lowest cost value or cost to perform. Moreover, the selected group of primitives 146 having the lowest cost value may then be selected as final cleanup pass profile 150 for the given work surface 122 and pass target 144. As FIG. 5 suggests, such cost value comparisons may be performed for each new variation of the cleanup pass profile 150 over a number of successive iterations until all available and feasible variations have been exhausted. In alternative embodiments, the controller 126 may first calculate all of the cost values for the original cleanup pass profile 150 and its variations, and subsequently determine the cleanup pass profile 150 with the lowest cost value in a single assessment.

Once the final cleanup pass profile 150 or the group of primitives 146 exhibiting the lowest cost value has been determined, the controller 126 may communicate the cost-optimized cleanup pass profile 150 to the machines 102 and/or machine implements 104 according to block 152-8 of FIG. 5. Specifically, the controller 126 may be configured to interrelate the polynomial or curve functions associated with the individual geometric primitives 146, such as via one or more mathematical relationships therebetween, in a manner which adjoins the endpoints 148 of the primitives 146. The controller 126 may further digitalize or otherwise translate the resulting functions defining the cleanup pass profile 150 into the appropriate instructions for execution by one or more of the machines 102 or machine implements 104 within the worksite 100. The controller 126 may communicate such instructions to the appropriate machines 102 and machine implements 104 via the communications device 130, or the like. Machines 102 or implements 104 receiving such instructions may then autonomously or semi-autonomously operate to execute the cleanup pass according to the cleanup pass profile 150.

From the foregoing, it will be appreciated that while only certain embodiments have been set forth for the purposes of illustration, alternatives and modifications will be apparent from the above description to those skilled in the art. These and other alternatives are considered equivalents and within the spirit and scope of this disclosure and the appended claims.

Claims

1. A computer-implemented method for determining a cleanup pass profile for a machine implement, comprising:

identifying a pass target extending from a first end to a second end along a work surface;

determining a plurality of primitives of a cleanup pass profile extending between the first end and the second end, each primitive being defined based at least partially on volume constraints;

determining a cost value associated with moving the machine implement along the cleanup pass profile based on an optimization cost function; and

adjusting the volume constraints associated with one or more of the primitives to minimize the cost value.

2. The computer-implemented method of claim 1, wherein the optimization cost function is configured to associate the cost value of the cleanup pass profile with one or more parameters including curvature, curvature limit, slope, slope limit, implement load, implement load average, maximum implement load limit, minimum implement load limit, crest volume, cut depth relative to pass target, cut depth relative to work surface, and volume differential between the pass target and the work surface.

3. The computer-implemented method of claim 2, wherein the optimization cost function is configured to increment the cost value for each parameter that exceeds a corresponding target threshold.

4. The computer-implemented method of claim 2, wherein each of the parameters is weighted by the optimization cost function according to its relevance to the cost value.

5. The computer-implemented method of claim 1, wherein the volume constraints associated with one or more of the primitives are adjusted using one or more of gradient descent algorithms, genetic algorithms, and Nelder-Mead algorithms.

6. The computer-implemented method of claim 1, wherein a total volume differential between the work surface and the cleanup pass profile is maintained during adjustments.

7. The computer-implemented method of claim 1, wherein an elevation and a slope of each endpoint of each primitive are maintained during adjustments.

8. The computer-implemented method of claim 1, further adjusting one or more of an elevation at one or more endpoints of the primitives, a location of one or more of the endpoints, and a number of the endpoints.

9. A control system for determining a cleanup pass profile for a machine implement, comprising:

a memory, the memory being a non-transitory computer-readable storage medium, configured to retrievably store one or more algorithms; and

a controller in communication with the memory and, based on the one or more algorithms, configured to at least:

identify a pass target extending from a first end to a second end along a work surface,

determine a plurality of primitives of a cleanup pass profile extending between the first end and the second end,

determine a cost value associated with moving the machine implement along the cleanup pass profile based on an optimization cost function, and

adjust volume constraints associated with one or more of the primitives to minimize the cost value.

10. The control system of claim 9, wherein the controller associates the cost value of the cleanup pass profile with one or more parameters of the optimization cost function including curvature, curvature limit, slope, slope limit, implement load, implement load average, maximum implement load limit, minimum implement load limit, crest volume, cut depth relative to pass target, cut depth relative to work surface, and volume differential between the pass target and the work surface.

11. The control system of claim 10, wherein the controller increments the cost value for each parameter of the optimization cost function that exceeds a corresponding target threshold.

12. The control system of claim 10, wherein the controller assigns a weight to each of the parameters of the optimization cost function according to its relevance to the cost value.

13. The control system of claim 9, wherein the controller adjusts the volume constraints associated with one or more of the primitives using one or more of gradient descent algorithms, genetic algorithms, and Nelder-Mead algorithms.

14. The control system of claim 9, wherein the controller maintains a total volume differential between the work surface and the cleanup pass profile during adjustments.

15. The control system of claim 9, wherein the controller maintains an elevation and a slope of each endpoint of each primitive during adjustments.

16. The control system of claim 9, wherein the controller adjusts one or more of an elevation at one or more endpoints of the primitives, a location of one or more of the endpoints, and a number of the endpoints.

17. (canceled)

18. (canceled)

19. (canceled)

20. (canceled)