Methods and control arrangements for operating autonomous agricultural vehicles

Abstract

This disclosure generally relates to livestock barn cleaning, and more specifically, to an autonomous agricultural vehicle comprising a cleaning device suitable for removing manure from livestock areas. According to a first aspect, this disclosure provides a method for operating the autonomous agricultural vehicle according to a cleaning plan. The method includes, while operating the autonomous agricultural vehicle according to the cleaning plan, monitoring at multiple individual times S1 a load representing the amount of material moved by the cleaning device, and adjusting the cleaning plan of the autonomous agricultural vehicle S5 based on changes in the load monitored at said individual times. This disclosure also relates to a method for operating the autonomous agricultural vehicle and a computer program for performing said method.

Description

Technical Field

[0001] This disclosure generally relates to livestock barn cleaning, and more specifically, to an autonomous agricultural vehicle comprising cleaning devices suitable for removing manure from livestock areas. This disclosure also relates to a method for operating the autonomous agricultural vehicle and a computer program for performing said method. Background Technology

[0002] An unavoidable consequence of livestock farming is the disposal of animal excrement, such as manure. The removal of animal excrement will be referred to as "barn cleaning" in this document. However, while animal excrement will be referred to as "manure" in this document, many other terms are used for excrement from different animal types, such as "effluent" or "feces." Automated barn cleaning offers a variety of solutions, such as various autonomous agricultural vehicles that move or collect manure, or scrapers typically pulled along channels by chains, ropes, or wires. What these solutions have in common is that they generally operate on predefined tracks according to a predefined schedule.

[0003] Autonomous agricultural vehicles are typically battery-powered. Optimization of travel distance per battery charge is often necessary to ensure optimal vehicle efficiency. WO 2011 / 062481 A1 discloses an autonomous agricultural vehicle for agricultural applications, comprising control circuitry to control a motor to adjust the speed of the autonomous vehicle in order to optimize battery efficiency. The control circuitry is configured to set the speed of the autonomous vehicle such that energy drawn from the battery is minimized over a predefined distance.

[0004] However, automated barn cleaning systems typically run their cleaning routines regardless of whether cleaning is needed, i.e., regardless of the amount of manure to be removed. This is considered inefficient. Summary of the Invention

[0005] The purpose of this disclosure is to mitigate at least some of the drawbacks of the prior art. Therefore, one objective is to provide a solution that enables more efficient livestock barn cleaning, as autonomous agricultural vehicles adapt their cleaning based on the actual time and location required for cleaning.

[0006] According to a first aspect, this disclosure proposes a method for operating an autonomous agricultural vehicle according to a cleaning plan, wherein the autonomous agricultural vehicle includes a cleaning device suitable for moving manure from a livestock area / livestock area. The method includes, while operating the autonomous agricultural vehicle according to the cleaning plan, monitoring a load representing the amount of material moved by the cleaning device at multiple individual times, and adjusting the cleaning plan of the autonomous agricultural vehicle based on changes in the load monitored at individual times. The method provides efficient cleaning because the cleaning volume will be based on a demand calculated from historical data. Furthermore, better energy optimization can be achieved because unnecessary cleaning is not performed.

[0007] In addition, maintenance and service requirements may be reduced as autonomous agricultural vehicles are used in a more efficient manner.

[0008] In some embodiments, a cleaning program defines a cleaning schedule for autonomous agricultural vehicles and / or one or more cleaning routes. This can also provide increased animal welfare and safety, as cleaning can be scheduled in times and places where animals are not present.

[0009] In some embodiments, the adjustment includes modifying the cleaning schedule to reduce changes in the monitored load. This allows for higher cleaning efficiency (in terms of the amount of material removed per unit time) and thus increases the cleaning capacity of autonomous agricultural vehicles.

[0010] In other words, embodiments of the subject matter described herein can be advantageously used to improve the efficiency of automated barn cleaning. For example, energy consumption can be minimized by cleaning only when and where it is needed, and by ensuring that the travel distance of the automated barn cleaning device is no greater than the absolute distance required to complete the work.

[0011] In some embodiments, the adjustments include modifying the cleaning schedule so that the monitored load remains below a predefined load threshold. This avoids excessively high loads on autonomous agricultural vehicles and cleaning devices. The result is typically less wear and tear on components.

[0012] In some embodiments, the change refers to a change over a period of time, and the adjustment of the cleaning schedule includes adjusting the cleaning schedule of the autonomous agricultural vehicle for a specific period of time based on changes over one or more previous periods. In some embodiments, the period of time is a weekday, a day, or 24 hours. In other words, the cleaning schedule is adapted based on historical periods of high material production (typically primarily manure). Thus, the user does not need to create the cleaning schedule, as this will be done by the proposed machine learning software. In other words, the autonomous agricultural vehicle will operate autonomously with little or no human intervention.

[0013] In some embodiments, adjusting the cleaning plan includes adjusting the cleaning schedule of autonomous agricultural vehicles such that, during a cleaning cycle, cleaning is performed more frequently during time intervals corresponding to a first average monitoring load than during time intervals corresponding to a second average monitoring load, where the first average monitoring load is higher than the second average monitoring load. In other words, cleaning is performed when it is deemed necessary.

[0014] In some embodiments, the method includes obtaining the corresponding location within the livestock area for each individual time period. In these embodiments, the variation represents a change in load at different locations within the livestock area, and adjustments to the cleaning schedule include adjusting the cleaning schedule of autonomous agricultural vehicles based on changes in load at different locations within the livestock area. In other words, the cleaning routes are adapted based on locations that historically produced large amounts of material (typically manure).

[0015] In some embodiments, adjusting the cleaning plan includes adjusting one or more cleaning routes and / or selecting one or more cleaning routes from a plurality of autonomous agricultural vehicle cleaning routes, such that cleaning is performed more frequently in areas corresponding to a first average monitoring load than in areas corresponding to a second average monitoring load, wherein the first average monitoring load is higher than the second average monitoring load. In other words, cleaning is performed where it is deemed necessary.

[0016] In some embodiments, the method includes calculating a material accumulation pattern representing the rate of material accumulation in a livestock area based on monitored load levels, and adjusting a cleaning schedule based on the calculated material accumulation pattern. In other words, it estimates the amount of material generated historically and adjusts the cleaning schedule accordingly. In some embodiments, the material accumulation pattern describes how material accumulation changes over time.

[0017] In some embodiments, the material accumulation pattern includes a material accumulation map that describes variations in material accumulation at different locations within the livestock area. Therefore, a multidimensional map representing material production within the livestock area can be used to enable the efficient adjustment of cleaning schedules.

[0018] In some embodiments, the method includes estimating a historical material accumulation rate in a livestock area based on monitored load levels, and predicting a future material accumulation rate in the livestock area based on the historical material accumulation rate. In these embodiments, adjustments include adjusting the cleaning plan based on the predicted future material accumulation rate in the livestock area. Thus, the amount of future material to be removed can be estimated in an accurate manner.

[0019] In some embodiments, the method includes obtaining user input indicating a desired level of cleanliness, and wherein adjustment includes adjusting the cleaning plan based on the obtained cleanliness level. Therefore, the user can set the desired cleanliness level based on a balance between cost and cleanliness.

[0020] In some embodiments, load capacity indicates the amount of energy required to move material accumulated over a period of time. In some embodiments, load capacity includes at least one of the following: power, pressure, weight, and flow rate.

[0021] In some embodiments, the method includes monitoring the load in livestock areas when it is believed that there is no material to be removed, while operating autonomous agricultural vehicles in the livestock areas, and calibrating the monitoring based on the monitored load indicating livestock areas without material. Therefore, the monitoring can be calibrated to account for constant (or slowly changing) conditions. In other words, a constant component of the load is not used when adjusting the cleaning schedule.

[0022] According to a second aspect, this disclosure provides a control arrangement for operating an autonomous agricultural vehicle configured to clean livestock areas according to a cleaning plan. The control arrangement is configured to monitor, at multiple individual times, a load representing the amount of material moved by the cleaning device while the autonomous agricultural vehicle is operating according to the cleaning plan, and to adjust the cleaning plan of the autonomous agricultural vehicle based on changes in the load monitored at the multiple individual times.

[0023] In some embodiments, the control arrangement is configured to perform a method according to one or more embodiments of the first aspect.

[0024] According to a third aspect, this disclosure proposes an autonomous agricultural vehicle configured to clean livestock areas according to a cleaning plan. The autonomous agricultural vehicle includes a motor, an energy storage device, a cleaning device, a load sensor, and a control arrangement according to the second aspect. The motor is configured to propel the autonomous agricultural vehicle. The energy storage device is configured to supply energy to the motor. The cleaning device is configured to remove dirt from the livestock area. The load sensor is configured to obtain a load indicating the amount of material moved by the cleaning device.

[0025] According to the fourth aspect, this disclosure relates to a computer program that includes instructions that, when a computer executes the program, cause the computer to perform the method according to the second aspect.

[0026] According to the fifth aspect, this disclosure relates to a computer-readable medium comprising instructions that, when executed by a computer, cause the computer to perform the method according to the second aspect. Attached Figure Description

[0027] Figure 1aAn autonomous agricultural vehicle according to some example embodiments is shown.

[0028] Figure 1b The concept of autonomous agricultural vehicles is shown in more detail.

[0029] Figure 2 A control arrangement according to some embodiments is shown.

[0030] Figure 3 The illustration shows a cleaning route for an autonomous agricultural vehicle in a livestock area according to an example embodiment.

[0031] Figure 4a and 4b This shows how much material is removed at each stage of the work when the cleaning schedule is adjusted.

[0032] Figure 5a and 5b The route of the autonomous agricultural vehicle in the livestock area is shown.

[0033] Figure 6 This is a flowchart of a method for operating autonomous agricultural vehicles.

[0034] Figure 7 The example material accumulation pattern is shown. Detailed Implementation

[0035] As previously mentioned, there are many different automated solutions for cleaning livestock sheds, such as dairy cow sheds. For example, autonomous farm vehicles have been developed for this purpose, which can be equipped with various navigation devices. For instance, induction wires or magnets can be fixed in the floor, and the autonomous vehicle can be equipped with sensors that allow the vehicle to be guided by the wires or magnets. Alternatively, navigation can be performed by beacons or repeaters placed along predetermined routes that can be detected by the autonomous device. The navigation device can thus guide the autonomous device along a predetermined route designed to cover as much of the shed as possible that needs cleaning at a given point. Cleaning schedules are typically pre-programmed based on time, causing the automated shed cleaning device to start and run at certain times. Therefore, conventional autonomous farm vehicles typically use fixed routes and fixed schedules to clean livestock areas. However, this cleaning method can be inefficient, given that the manure load in the shed can vary over time and varies from area to area, depending on, for example, how the animals move around.

[0036] This disclosure introduces an adaptive barn cleaning concept in which cleaning is performed where and when it is required. Some embodiments of the proposed technology are based on the insight that animal behavior typically does not change drastically from day to day. Therefore, optimal cleaning can often be achieved by studying animal behavior over the preceding days and then adjusting the cleaning accordingly. More specifically, this paper proposes monitoring the amount of material removed by the cleaning apparatus of an operated autonomous agricultural vehicle during barn cleaning and using monitoring data representing historical amounts of material removed to operate the autonomous agricultural vehicle in a more efficient manner.

[0037] More specifically, this paper proposes measuring the amount of material removed by autonomous agricultural vehicles during barn cleaning, for example, by monitoring motor load and battery voltage. Based on the monitored values, software can be used to optimize cleaning schedules to clean more frequently when high loads are recorded during the day and less frequently when they are not. Load measurement also makes it possible to identify animal traffic patterns and use them to avoid cleaning in crowded areas to avoid disturbing the animals (e.g., during feeding times, in waiting areas, etc.).

[0038] Furthermore, the proposed solution allows farmers to configure cleanliness levels. Therefore, farmers will be able to make trade-offs between cleanliness levels and travel distance per battery charge.

[0039] The following will describe exemplary embodiments of autonomous agricultural vehicles and methods for operating such vehicles with reference to Figures 1-6. The embodiments are intended to be performed by a control arrangement associated with the autonomous agricultural vehicle, which includes cleaning devices suitable for removing manure from livestock areas. The term "associated with" is intended herein to at least cover a control arrangement operatively connected to the autonomous agricultural vehicle. The following will be combined with… Figure 3 A more detailed description of the control layout is provided. Autonomous agricultural vehicle systems are suitable for automatically cleaning barns for dairy animals such as cows, buffalo, sheep, or goats.

[0040] Figure 1a An autonomous agricultural vehicle 1 is shown, which includes a cleaning device 13 suitable for removing manure from the livestock area 3. Figure 3 The autonomous agricultural vehicle 1 shown is a cleaning robot that can be controlled to navigate and move relatively freely in at least two physical dimensions, i.e., not limited to a fixed track. Therefore, the autonomous agricultural vehicle 1 is configured to move autonomously around livestock buildings or in fields.

[0041] Figure 1b This is a conceptual diagram of the components of an autonomous agricultural vehicle 1 and a user device 2. The autonomous agricultural vehicle 1 shown includes a motor 11, energy storage 12, a cleaning device 13, a load sensor 14, and a control arrangement 10. It must be understood that the autonomous agricultural vehicle 1 includes... Figure 1bOther components not shown include those for steering, braking, and charging of the autonomous agricultural vehicle 1, as well as sensors for autonomous control. However, for simplicity, this document only describes components relevant to the proposed technology.

[0042] Motor 11 is configured to propel autonomous agricultural vehicle 1. More specifically, motor 11 is configured to convert energy supplied by energy storage device 12 into mechanical force. Motor 11 is, for example, an electric motor. Energy storage device 12 is configured to supply energy to motor 11. Energy storage device 12 is, for example, a battery. Energy storage device 12 is typically connected to docking station 34 ( Figure 3 The autonomous agricultural vehicle 1 is parked there between cleaning work phases while charging.

[0043] The cleaning device 13 is a device configured to remove dirt from livestock areas. In the illustrated embodiment, the cleaning device is a scraper disposed on the front of the autonomous agricultural vehicle 1. In some embodiments, the autonomous agricultural vehicle 1 operates in a livestock area with a slatted floor, whereby the scraper scrapes material (e.g., manure) downwards into the holes of the slatted floor. Alternatively, the material can be dumped into a ditch along a cleaning path.

[0044] In some embodiments, the cleaning device 13 includes a material pickup arrangement, such as a suction arrangement (e.g., a vacuum cleaner arranged to suction material upwards) or a pickup belt. In other words, in some embodiments, the autonomous agricultural vehicle includes a container, such as a tank, and a material pickup device arranged to move material into the tank. Such a pickup device may be used in conjunction with a scraper arranged to collect material at the inlet of the pickup arrangement and / or with a slatted floor. The cleaning device 13 may also, or alternatively, include a rotary brush, a water sprayer, and / or any other suitable cleaning device or combination thereof.

[0045] Load sensor 14 is configured to obtain a load indicating the amount of material removed by cleaning device 13. In some embodiments, load sensor 14 is a power sensor arranged to measure the power (or energy) required to remove the material. For example, the load sensor may be arranged to sense the power required to propel the autonomous agricultural vehicle 1. Alternatively, load sensor 14 may be arranged to measure the load on a material pickup device, such as a suction device or a conveyor belt. In other words, load sensor 14 is configured to measure the load caused by the removal of material. The load required to propel the material can be estimated by subtracting the power required to propel the autonomous agricultural vehicle 1 on a clean floor from the actual power required. In other words, the amount of material removed by the autonomous agricultural vehicle 1 can be estimated by monitoring the motor power while operating the autonomous agricultural vehicle 1.

[0046] If the autonomous agricultural vehicle 1 includes a container (e.g., a tank), the amount removed can be estimated by measuring the volume and / or weight of the material accumulated in the container. In other words, in some embodiments, the load sensor includes a weight sensor configured to measure the amount of weight and / or a volume sensor configured to measure the amount of volume. In some embodiments, the load sensor is a flow sensor arranged to measure the amount of flow representing the flow rate of material entering the container.

[0047] In some embodiments, the load sensor is a force sensor arranged to measure force. The measured pressure may correspond, for example, to the force required to move material. For example, pressure on a scraper or brush may be measured. Typically, the load sensor may include any type of sensor or combination of sensors adapted to obtain an estimate of the amount of material removed by the cleaning device during operation. In some embodiments, the load sensor includes a battery charge monitor. In some embodiments, the load sensor is an instrument that receives data from a motor drive or any other device to be configured to monitor the load required to remove material (directly or indirectly), suitable for monitoring the amount of material moved by the autonomous agricultural vehicle 1.

[0048] The control arrangement 10 is configured to autonomously operate the autonomous agricultural vehicle 1 along a cleaning route. This typically involves the propulsion, braking, and steering of the autonomous agricultural vehicle 1. In some embodiments, it also involves tracking the position of the autonomous agricultural vehicle 1. Position tracking can be achieved using computation combining data from various sensors, such as optical sensors, wireless sensors, etc. The control arrangement can also be configured to detect obstacles (e.g., animals) in the cleaning route and control the autonomous agricultural vehicle 1 to avoid such obstacles. Specifically, the control arrangement 10 is configured to control the motor 11 to propel the autonomous agricultural vehicle 1.

[0049] Figure 2A control arrangement 10 according to some embodiments is shown in more detail. The control arrangement 10 includes hardware and software. For example, the hardware consists of various electronic components, such as those on a printed circuit board (PCB). The most important of these components are typically a processor 101, such as a microprocessor, and a memory 102, such as an EPROM or flash memory chip. The software is typically software code running in a microcontroller. The control arrangement 10 will be described in more detail below. The control arrangement 10 shown also includes a communication interface 103. The communication interface 103 is configured for signal and / or data communication between the control arrangement 103 and a remote device, such as a user device 2. The communication interface 103 is configured for wireless communication using any suitable protocol, such as Bluetooth or IEEE 802.11. The communication interface 103 can also be configured for wired communication, for example, via a docking station. In some embodiments, the control arrangement 10 is a functional device. Therefore, the control arrangement 10 can be distributed among multiple physical control units, some of which may be located outside the autonomous agricultural vehicle 1.

[0050] In some embodiments, the autonomous agricultural vehicle 1 is configured to communicate with a user device 2 using a communication interface 103. A user can use the user device 2 to input user information for use by the autonomous agricultural vehicle 1. Information can also be provided to another user device 2 using the user device 2.

[0051] In the illustrated example, the user device is a smartphone. In other embodiments, the user device is a laptop computer, tablet computer, or any other device. User device 2 includes a control arrangement 20 and a display 21, which is here a touch display capable of presenting a graphical user interface. In other embodiments, the user device may include other means for receiving user input and providing information to the user, such as buttons. In some embodiments, the user device includes a software application configured to perform portions of the methods presented herein.

[0052] Control arrangement 20 includes hardware and software. For example, the hardware consists of various electronic components on a printed circuit board (PCB). The most important of these components are typically a processor 101, such as a microprocessor, and a memory 102, such as an EPROM or flash memory chip. The software is typically software code running in a microcontroller. Control arrangement 20 also includes a communication interface configured to enable communication with the control arrangement 103 of the autonomous agricultural vehicle 1.

[0053] Figure 3 The cleaning route 31 of an autonomous agricultural vehicle 1 in a livestock area 30 is shown according to an example embodiment. Figure 3A schematic top view of the livestock area 30 is shown, in which the cleaning route 31 of the autonomous farm vehicle 1 is indicated by a dashed line. The autonomous farm vehicle 1 is arranged to operate in the aisles 32 (shaded area) of the livestock area. Along both sides of the aisles 32 are pens 33 where animals can stay, be tethered, or untied. For example, the livestock area 30 is configured for dairy cows. A docking station 34 for the autonomous farm vehicle 1 can be accessed from the aisles 32.

[0054] The proposed solution makes it possible to adjust the cleaning schedule of the autonomous agricultural vehicle 1 based on the amount of material removed by the autonomous agricultural vehicle 1 in previous cleaning phases. This can be accomplished, for example, by adjusting the cleaning schedule or cleaning route of the autonomous agricultural vehicle 1.

[0055] Figure 4a and 4b This illustration shows how the amount of material removed in each work phase is affected when adjusting a cleaning schedule according to some embodiments of the present disclosure. A cleaning schedule is a process that defines the sequence (i.e., start time) of cleaning work phases over a period of time (e.g., a day). The amount of material (primarily manure) produced in a livestock area varies at different times of the day. Therefore, if an autonomous agricultural vehicle 1 is programmed to perform cleaning work phases periodically (in cycles Δt) along a predefined cleaning route throughout the day, the total amount of material removed Δmat in each cleaning work phase 41 will vary, such as... Figure 4a As shown.

[0056] Experiments show that the amount of material produced varies daily. Nevertheless, each day exhibits a similar trend specific to the particular livestock area. Therefore, data from the past few days can be used to predict material production. By using a cleaning schedule based on such predictions, cleaning phases can be scheduled as needed, rather than using fixed cycles.

[0057] Figure 4b This graph shows that, compared to using a periodic schedule, if the cycle is adjusted based on predicted manure volume, the total amount of material removed Δmat in each cleaning phase 41 will vary less within the cleaning phase. It is clear that, compared to... Figure 4a In Figure 4b In the example, the difference between the maximum and minimum amount of material removed during a cleaning work phase, vmat (i.e., the change in the amount of material removed), was significantly reduced.

[0058] Figure 5a and 5bThe diagram illustrates the route of an autonomous agricultural vehicle within a livestock area. The amount of material (primarily manure) produced within the livestock area will vary between different sections. For example, one section 51 of the livestock area may contain more animals than the rest of the livestock area 30. Other sections may even be empty. Alternatively, the amount of material produced may vary for other reasons, such as how the animals move. Therefore, during each cleaning phase, along... Figure 3 The entire cleaning route 31 shown may not be efficient for operating the autonomous agricultural vehicle 1. In other words, the cleaning route of the autonomous agricultural vehicle 1 can be adjusted. For example, a shorter cleaning route 52 can be defined. For example, a shortened route can be used every two cleaning work phases. Alternatively, several different routes (or loops) can be defined, and the cleaning plan can then include a selected combination of loops based on the predicted amount of material produced in the livestock area at different times.

[0059] Reference Figure 6 Flowcharts and Figure 2 The proposed technology is described in more detail for livestock areas. Figure 5 illustrates an exemplary method of operating autonomous agricultural vehicle 1 according to a cleanup plan. The method in Figure 5 is, for example, performed by autonomous agricultural vehicle 1 ( Figure 1a ) control layout 10 ( Figure 1b The method can be implemented as a computer program comprising instructions that are executed in a computer (e.g., a processor in a control arrangement). Figure 2 When the program is executed, it causes the computer to perform the method. According to some embodiments, the computer program is stored in a computer-readable medium (e.g., memory or compressed optical disc) that includes instructions that, when executed by the computer, cause the computer to perform the method.

[0060] The method is typically performed continuously or repeatedly during the operation of autonomous agricultural vehicle 1. Autonomous agricultural vehicle 1 operates according to a cleaning plan that includes a cleaning schedule and a cleaning route 31. The cleaning route defines where autonomous agricultural vehicle 1 should travel, and the cleaning schedule defines when autonomous agricultural vehicle 1 should travel. In other words, the cleaning plan defines the sequence of cleaning work phases for autonomous agricultural vehicle 1. When autonomous agricultural vehicle 1 is not cleaning, it is typically at docking station 34 ( Figure 3 Charging in the middle.

[0061] When the autonomous agricultural vehicle 1 is first used in the livestock area 30, it must be programmed with an initial cleaning plan. As previously mentioned, such a cleaning plan typically includes a cleaning schedule and a cleaning route. In some embodiments, the cleaning schedule is pre-programmed by the manufacturer. Alternatively, the user must configure the cleaning schedule. However, the initial cleaning route, which includes one or more possible loops, typically needs to be programmed at installation. A loop may, for example, cover the entire livestock area 30 and is typically used in the initial cleaning plan. However, alternative loops that only cover portions of the livestock area 30 can also be configured. In other words, in some embodiments, the method includes configuring the S0 autonomous agricultural vehicle 1 to operate according to the initial cleaning plan.

[0062] Initially, the method may also include user input to obtain the desired cleanliness level of S00, wherein the desired cleanliness level may, for example, include some predefined levels, such as level 1, level 2, and level 3.

[0063] User input is provided to the control arrangement 10, for example, via communication interface 103. Communication interface 103 then includes or communicates with a user interface (not shown). In some embodiments, the user interface may be presented on a user device, such as on a display of the user device. Figure 1b The user interface may include input devices such as a touchscreen, keyboard, or microphone.

[0064] The suggested cleanliness level can be communicated via a user interface, such as through a display. The suggested cleanliness level could be, for example, a balanced cleanliness level, striking a balance between cleaning costs and increased cleanliness. The user can then select the balanced cleanliness level, or decrease or increase the cleanliness within certain limits. Therefore, user input can be one or more touches on a touchscreen, one or more keyboard inputs, or voice commands.

[0065] The autonomous farm vehicle 1 is then operated using the initial cleaning schedule. Therefore, the cleaning route for each work phase is typically the same (covering the entire livestock area), and consequently, the time between each cleaning work phase is usually always the same, such as... Figure 4a As shown. During these initial cleaning phases, training data is collected to adjust the cleaning schedule. In other words, the method includes monitoring the load S1a, representing the amount of material moved by the cleaning device 13, at multiple individual times while operating the autonomous agricultural vehicle 1 according to the initial cleaning plan. Monitoring includes using a load sensor 14 ( Figure 1b Sensing load quantity. Load quantity can be combined with... Figure 1bAny of the described load amounts. Therefore, the amount of material moved by the cleaning device 13 is being measured. In other words, information about animal behavior in the livestock area is collected during these cleaning work phases. Monitoring S1 may include continuous monitoring of the load while operating the autonomous agricultural vehicle 1, or monitoring once or several times during each cleaning work phase. Typically, the load is monitored during several consecutive cleaning work phases. In some embodiments, monitoring S1 is performed during each cleaning work phase performed in the livestock area 30. If the cleaning route is along a lane where there is one (or more) locations where material can be dumped using, for example, a scraper, it may be sufficient to measure the load just before dumping the material, i.e., when the amount of material in the scraper 1 is at its maximum and just before the material leaves the scraper.

[0066] Once training data has been collected, the S5 cleaning plan can be adjusted based on the monitored load, as will be explained further below. However, monitoring S1 will proceed in the same manner regardless, the only difference being that the cleaning plan is different, i.e., it is adjusted. In other words, the method then includes monitoring the load of material moved by the cleaning device 13 at multiple individual times, represented by S1b, while operating the autonomous agricultural vehicle 1 according to the adjusted cleaning plan.

[0067] It may be desirable to correlate the monitored load with the corresponding location of the autonomous agricultural vehicle 1. Therefore, in some embodiments, the method includes obtaining the corresponding location within the livestock area 30 for each individual time period. The autonomous agricultural vehicle 1 must typically know its location within the livestock area 30 in order to navigate correctly according to the cleaning route and also avoid obstacles. Therefore, the current location is usually already available in the control arrangement 10, or at least can be calculated from already available information.

[0068] Monitoring S1 may also include storing the value of the sensed load and the corresponding location (if obtained) in a memory such as the memory 102 of the control arrangement 10, or in a remote storage device referred to as the “cloud”.

[0069] The monitored load indicates how much material has been removed by the autonomous farm vehicle 1. By knowing how much material has been removed, it is also possible to estimate how much material has been generated since the last cleaning phase. This “material generation rate” is referred to herein as the material accumulation rate. To obtain an accurate estimate of this accumulation rate, it is generally necessary to derive an offset representing the constant conditions associated with livestock area 30 and autonomous farm vehicle 1, as will be explained further below.

[0070] The proposed technique is based on the insight that the material accumulation rate varies over time and also differs between different areas within livestock area 30. Patterns of these variations can be determined by studying monitored load levels. These patterns are referred to herein as material accumulation patterns. In other words, the patterns visualize different trends in material accumulation, which can be determined based on historically monitored load levels. Therefore, in some embodiments, the method includes calculating S4, representing the material accumulation pattern of the material accumulation rate in livestock area 30, based on monitored load levels. This pattern typically shows how material accumulation in livestock varies over a period of time, such as a day, and between different parts of the day. In other words, machine learning algorithms are used to predict material accumulation using data from the past few days. The patterns can identify trends within a day as well as trends in specific parts of the livestock area.

[0071] A material accumulation pattern illustrates how the material accumulation rate changes over time. In other words, in some embodiments, the material accumulation pattern describes how material accumulation changes over time. This pattern shows how the material accumulation of livestock changes over a period of time, such as during a day. For example, it may be higher at some individual times of the day and lower at some individual times. In a simple example, the material accumulation rate across the entire livestock area can be graphically represented as a function of time. The accumulation rate between the two cleaning stages can be estimated by dividing the total amount of material removed in the second cleaning stage by the time elapsed since the first cleaning stage.

[0072] In some embodiments, a pattern for the following day is created by studying the average accumulation at different times over the previous few days. For example, a moving average is used.

[0073] Material accumulation patterns illustrate how the material accumulation rate varies spatially. In other words, in some embodiments, the material accumulation pattern includes a material accumulation map that describes the variation in material accumulation at different locations within a livestock area. Therefore, the pattern can illustrate how accumulation occurs in different parts of the livestock area. This can be illustrated on a map of the barn, where different patterns can indicate different amounts of material, such as those used in [the specific example]. Figure 7 As illustrated in the examples.

[0074] Of course, patterns representing location and time can also be estimated. Such patterns would include multiple dimensions and can be represented as a map of livestock area 30.

[0075] The method further includes adjusting the cleaning schedule of the S5 autonomous agricultural vehicle 1 based on changes in load monitored over individual time periods. In other words, the cleaning schedule is adjusted to compensate for changes in the amount of material generated. In this way, cleaning may be more efficient.

[0076] In some embodiments, the variation is the variation between different parts of the livestock area 30. Thus, the variation is the variation between different points in time within one (or more) individual cleaning work phases. For example, the average amount of material removed at location A is compared to the average amount of material removed at location B.

[0077] In some embodiments, the variation is the variation between different times, such as different times of day. So, the variation is the variation between different points in time during different cleaning work phases. For example, comparing the average amount of material removed during a cleaning work phase performed at 8:00 AM with the average amount of material removed during a cleaning work phase performed at 11:00 AM.

[0078] As previously mentioned, the objective is typically to balance the differences in the amount of material removed during each cleaning phase. In other words, in some embodiments, adjustment S5 includes adjusting the cleaning schedule to reduce variations in the monitored load. This is typically achieved by performing cleaning at the required times and places. In other words, in some embodiments, adjusting the cleaning schedule includes adjusting the cleaning timetable of the autonomous agricultural vehicle 1 so that, during a cleaning cycle, cleaning is performed more frequently during time intervals corresponding to a first average monitored load than during time intervals corresponding to a second average monitored load, where the first average monitored load is higher than the second average monitored load. In other words, adjustment S5 aims to schedule more cleaning when material accumulation is historically high.

[0079] In some embodiments, adjusting S5 includes adjusting the cleaning plan based on an obtained cleanliness level, such as a cleanliness level input by the user. The desired cleanliness level may define the maximum and / or minimum number of cleaning work stages for each time period. Alternatively, the desired cleanliness level may define the maximum and / or minimum distance traveled during cleaning in each time period.

[0080] In some embodiments, adjusting the cleaning plan includes adjusting one or more cleaning routes of the autonomous agricultural vehicle 1. This means performing cleaning more frequently in areas corresponding to a first average monitoring load than in areas corresponding to a second average monitoring load. The first average monitoring load is higher than the second average monitoring load. In other words, adjustment S5 should strive to schedule more cleaning in locations with historically high material accumulation. This can be achieved by modifying the last used cleaning route. Alternatively, this can be achieved by pre-configuring multiple alternative cleaning routes (e.g., different loops) and dynamically selecting one or more of them for each specific cleaning work phase.

[0081] If a material accumulation pattern has been calculated, adjusting S5 involves adjusting the cleaning schedule based on the calculated pattern. For example, if the pattern indicates higher material output at certain locations and / or times, the cleaning schedule can be adjusted to perform more cleaning more frequently at those locations and / or times. Figure 5a -b is shown.

[0082] In some embodiments, adjusting S5 includes adjusting the cleaning schedule so that the monitored load remains below a predefined load threshold. The predefined load threshold may be fixed or configurable. For example, it may be based on a cleanliness level selected by the user, as explained above.

[0083] This can be achieved in different ways. For example, when a large amount of material is expected to be generated, cleaning can be performed more frequently, such as... Figure 4a -b is shown. If the cleaning device 13 is a scraper, it means adjusting the cleaning schedule so that the estimated amount of material accumulating in the scraper does not exceed a predefined level, because a large amount of material would cause a high load on the motor 11. To avoid this, for example, the path can be adjusted so that the scraper is emptied more frequently and / or cleaning work phases are scheduled more frequently. This is especially relevant when material is collected in the container of the autonomous agricultural vehicle 1, which may otherwise be too full.

[0084] If the change represents a change over a certain time period, then the change over said time period can be analyzed. Typically, this type of change is determined by averaging over several previous time periods. For example, if the load in the morning is always higher than in the afternoon, the cleaning schedule can be adjusted accordingly. In other words, in some embodiments, the adjustment S5 of the cleaning plan includes adjusting the cleaning schedule of the autonomous agricultural vehicle 1 for a specific time period based on changes over one or more previous time periods.

[0085] If the location of the autonomous agricultural vehicle 1 is obtained, the variation represents the change in load at different locations within the livestock area 30. Typically, this type of variation is determined by averaging over several previous cleaning phases. For example, if the load in one section 51 (Figure 5) of the livestock area 30 is consistently higher than in the rest of the livestock area 30, the cleaning route can be adjusted so that more cleaning is performed in that section 51 (Figure 5). This can be done by adjusting the cleaning route so that the shorter cleaning route 51 is performed once every two cleaning operations and the entire cleaning route 31 (Figure 5) is performed more frequently. Figure 3 This is performed every two times. In other words, in some embodiments, the adjustment of the cleaning plan S5 includes adjusting the cleaning schedule of the autonomous agricultural vehicle 1 based on changes in load at different locations in the livestock area.

[0086] To efficiently calculate the amount of material removed by cleaning device 13, calibration may be required. This may include monitoring the load when no material is being removed. This can be performed by operating autonomous farm vehicle 1 when livestock area 30 is clean (i.e., material in free form). The offset of the load, which is not caused by the removed material but by other factors such as friction, is then estimated. This offset load is typically required to correctly estimate the amount of material removed by autonomous farm vehicle 1. More specifically, the offset must typically be subtracted from the monitored load to correctly estimate the removed material. In other words, in some embodiments, the method includes calibrating monitoring S1 based on the monitored load representing livestock areas without material.

[0087] This operation can be completed before the autonomous farm vehicle 1 is put into use. However, for good results, the floor in the livestock area 30 should already be covered with some layer of manure; after some animals are present. Otherwise, the autonomous farm vehicle 1 will be driving on bare concrete, which may yield incorrect results. It can also be repeated from time to time after the autonomous farm vehicle 1 is put into use, as the offset may change due to, for example, wear and tear of hardware components or changes in the foundation. In other words, in some embodiments, the method includes monitoring the load of S1c while operating the autonomous farm vehicle 1 in the livestock area when it is believed that there is no material to be removed from the livestock area 30. Such operations are performed, for example, at night when the animals are not disturbed.

[0088] In a non-limiting illustrative example, an operator wants to operate an autonomous agricultural vehicle 1 in livestock area 30. The operator needs a balanced cleanliness level, which is a trade-off between cleanliness and power consumption. The operator inputs the balanced cleanliness level to control arrangement 10 via a remote interface. Control arrangement 10 thus receives user input (S00) indicating the desired cleanliness level.

[0089] The user also configured the initial cleaning route for autonomous agricultural vehicle 1. In this example, the initial cleaning route is... Figure 3 Route 31, as defined in the map, covers all the lanes of the livestock area 20. When configuring the route, other map data related to the livestock area 30 is also entered, giving the control arrangement 10 information about the places it can travel within the livestock area. Therefore, the control arrangement 10 is configured with information about alternative routes that can be used.

[0090] In this example, control arrangement 10 selects an initial cleaning schedule based on the input cleanliness level. This initial cleaning schedule comprises nine individual cleaning work phases, each with equal time intervals. The initial cleaning schedule spans one day. The cleaning work phases are referred to herein as Phase 1, Phase 2, ..., Phase 9.

[0091] In other words, autonomous agricultural vehicle 1 is now configured to operate according to the initial cleaning plan (S0). Then, autonomous agricultural vehicle 1 operates according to the initial cleaning plan over a multi-day period. In this example, the initial cleaning plan is executed for four days. Therefore, nine cleaning work phases (phases 1-9) are performed each day.

[0092] A load sensor 14 is previously installed within a motor controller (not shown) to control motor 11 to propel the autonomous agricultural vehicle 1 (without interfering with the control signal). This load sensor monitors the load on motor 11, here by measuring the power supplied to motor 11. As explained above, this load indicates the current amount of material being pushed by the cleaning device 13 (e.g., a scraper) of the autonomous agricultural vehicle 1. During cleaning according to the initial cleaning plan, the load amount S1a is continuously monitored at various time points. For each monitored load amount, the corresponding position S2 is also obtained. For each position where the load amount is monitored, the corresponding material accumulation rate can be estimated by studying how much the load amount has increased since the previous load amount was obtained and dividing this increase by the time elapsed since the previous load amount was obtained. Typically, observing the load amount is most relevant after inferring an offset indicating "no material removal," as discussed above. If the scraper is cleared in between, the accumulation of the previous load amount is naturally zero.

[0093] Data is stored in the cloud. Control arrangement 10 is therefore configured to communicate with the remote storage device, referred to herein as the "cloud." The S3 material accumulation pattern is then calculated using data collected during the initial four days of cleaning work phases (1-9). In this example, the material accumulation rate representing a specific location and a specific work phase number is calculated by averaging the material accumulation rates for each location and work phase number over the past four days. Figure 7 In the example, the resulting patterns are presented in the form of maps (one map for each cleaning phase 1-9), where different patterns are used to represent the material accumulation rate at different locations in the barn, and the density of the pattern corresponds to the material accumulation rate ("low", "normal", "high").

[0094] The estimated material accumulation pattern over the past four days is then used to predict the material accumulation rate of S4 in the livestock area for the following day, i.e., the future material accumulation rate. In this example, it can be noted that the material accumulation rate before cleaning work stage 1 and between cleaning work stages 1 and 2 is very low. For example, it is below the predefined threshold indicating "low accumulation". It can be noted that the material accumulation rate between cleaning work stages 4 and 5 and between cleaning work stages 6 and 7 is very high. For example, it is above the predefined threshold indicating "high accumulation". During cleaning work stages 6 and 7, it is only excessively high in one part 32 of the livestock area. Between other cleaning work stages, material accumulation is considered "normal". Note that these thresholds may differ for different cleanliness levels.

[0095] Based on these findings, the cleaning plan is then adjusted (S5). Therefore, in this example, the adjustment includes omitting work stage 1, since all material expected to accumulate before work stage 2 can be removed in a single cleaning stage. Furthermore, an additional cleaning stage can be inserted between cleaning stages 4 and 5 to avoid high loads in work stage 5. Additionally, an extra cleaning stage can be inserted between cleaning stages 6 and 7, but only covering the portion of the livestock area where the material accumulation rate exceeds the threshold indicating "high accumulation."

[0096] The following day, autonomous agricultural vehicle 1 was operated according to the adjusted cleaning plan. Then, during the cleaning work phases performed according to the updated cleaning schedule, the load of S2b was monitored. Subsequently, steps S3-S5 were repeated using the monitored load from the cleaning work phases performed during the four most recent cleaning schedules. Therefore, data from the three most recently executed time periods when using the initial cleaning plan was added to the data from the last time period when using the adjusted cleaning schedule.

[0097] Therefore, the cleaning schedule can be adjusted based on changes in daily monitoring data. The method can be executed more or less automatically. In some embodiments, the desired level of cleanliness can be fixed or predetermined. The method can be executed fully automatically. Then, the user manually creates any cleaning schedule. Therefore, the cleaning process can be optimized for one or more of the following: power consumption, machine wear and tear, maintenance and service time, animal welfare, and safety.

[0098] It must be anticipated that the illustrative examples above are simplified. In actual implementation, the algorithms may become more complex, employing, for example, machine learning algorithms. For instance, the timing of several or all cleaning phases might be slightly adjusted. Additionally, the cleaning routes might be adjusted for several or all cleaning phases. For example, short segments of the route might be skipped or added to adapt the cleaning to predicted material buildup in livestock areas.

[0099] This disclosure also relates to a corresponding control arrangement 10 for operating an autonomous agricultural vehicle 1, the autonomous agricultural vehicle 1 being configured to clean a livestock area 30 according to a cleaning plan, see [link to relevant documentation]. Figure 2 More specifically, the control arrangement 40 is configured to monitor the load amount, representing the amount of material moved by the cleaning device 13, at multiple individual times while the autonomous agricultural vehicle 1 is operating according to the cleaning plan, and to adjust the cleaning plan of the autonomous agricultural vehicle 1 based on the changes in the load amount monitored at the multiple individual times.

[0100] Control arrangement 40, or more specifically, processor 41 of control arrangement 40, is configured to cause control arrangement 40 to perform all aspects of the methods described in FIG. 5. This is typically accomplished by running computer program code stored in memory 42 in processor 41 of control arrangement 40.

[0101] The terminology used in the description of the embodiments illustrated in the accompanying drawings is not intended to limit the described methods, control arrangements, or computer programs. Various changes, substitutions, and / or modifications may be made without departing from the disclosed embodiments as defined by the appended claims.

[0102] As used herein, the term “or” should be interpreted as mathematical OR, i.e., inclusive disjunction; unless otherwise expressly stated, it cannot be interpreted as mathematical XOR. Furthermore, unless otherwise expressly stated, the singular forms “a,” “an,” and “the” will be interpreted as “at least one,” and thus may include multiple entities of the same kind. It should be further understood that the terms “includes,” “comprises,” “including,” and / or “comprising” specify the presence of the stated features, actions, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, actions, integrals, steps, operations, elements, components, and / or groups thereof. A single unit, such as a processor, can implement the functions of several items recited in the claims.

Claims

1. A method for operating an autonomous agricultural vehicle (1) according to a clean-up plan, wherein, The cleaning plan defines a cleaning schedule and / or one or more cleaning routes (31) for the autonomous agricultural vehicle (1), the autonomous agricultural vehicle (1) including a manure cleaning device (13) suitable for moving livestock areas (30) and a load sensor (14), the method comprising: - While operating the autonomous agricultural vehicle (1) according to the cleaning plan, the load amount, representing the amount of material moved by the cleaning device (13), is monitored at multiple individual times via the load sensor, and - The cleaning plan of the autonomous agricultural vehicle (1) is adjusted by adjusting the cleaning schedule of the autonomous agricultural vehicle (1) and / or one or more cleaning routes (31) based on the changes in load monitored during the individual time periods.

2. The method of claim 1, wherein the adjustment includes adjusting the cleaning schedule such that the change in the monitored load is reduced.

3. The method of claim 1, wherein the adjustment includes adjusting the cleaning schedule such that the monitored load remains below a predefined load threshold.

4. The method of claim 1, wherein the change represents a change over a period of time, and wherein the adjustment of the cleaning plan includes adjusting the cleaning schedule of the autonomous agricultural vehicle (1) for the period of time based on changes over one or more previous periods of time.

5. The method according to claim 4, wherein the time period is a workday, a day, or 24 hours.

6. The method of claim 4, wherein the adjustment of the cleaning schedule comprises adjusting the cleaning schedule of the autonomous agricultural vehicle (1) such that cleaning is performed more frequently during a cleaning cycle in a time interval corresponding to a first average monitoring load than in a time interval corresponding to a second average monitoring load, wherein the first average monitoring load is higher than the second average monitoring load.

7. The method according to claim 1, comprising: - For each individual time period, obtain the corresponding location within the livestock area. And the change refers to the change in the load at different locations within the livestock area, and The adjustment of the cleaning plan includes adjusting the cleaning schedule of the autonomous agricultural vehicle (1) based on the changes in the load at different locations in the livestock area.

8. The method of claim 7, wherein the adjustment of the cleaning plan includes adjusting the cleaning route of the autonomous agricultural vehicle (1) such that cleaning is performed more frequently in areas corresponding to a first average monitoring load than in areas corresponding to a second average monitoring load, wherein the first average monitoring load is higher than the second average monitoring load.

9. The method according to claim 1, comprising: -Based on the monitored load, calculate the material accumulation pattern representing the material accumulation rate in the livestock area (30), and The adjustments mentioned therein include adjusting the cleaning plan based on calculated material accumulation patterns.

10. The method of claim 9, wherein the material accumulation pattern describes the change of material accumulation over time.

11. The method of claim 9 or 10, wherein the material accumulation pattern includes a load accumulation map that describes the variation of material accumulation at different locations within the livestock area.

12. The method of claim 1, further comprising monitoring the load while operating the autonomous agricultural vehicle (1) in the livestock area when it is determined that there is no material to be removed in the livestock area, and calibrating the monitoring based on the monitored load indicating that there is no material in the livestock area.

13. The method of claim 1, further comprising: - Based on the monitored load, estimate the historical material accumulation rate in the livestock area, and - Based on the historical material accumulation rate, predict the future material accumulation rate in the livestock area. The adjustments include adjusting the cleaning plan based on the projected future material accumulation rate in the livestock area.

14. The method of claim 1, wherein the method comprises: - Obtain user input indicating the desired level of cleanliness, and wherein the adjustment includes adjusting the cleaning plan based on the obtained level of cleanliness.

15. The method of claim 1, wherein the load indicates the amount of energy required to move the accumulated material during a certain time period.

16. The method of claim 1, wherein the load includes at least one of the following: power, pressure, weight, and flow rate.

17. A control arrangement for operating an autonomous agricultural vehicle (1), said autonomous agricultural vehicle (1) being configured to clean livestock areas according to a cleaning plan, wherein, The cleaning plan defines the cleaning schedule and / or one or more cleaning routes (31) of the autonomous agricultural vehicle (1), the autonomous agricultural vehicle (1) including a manure cleaning device (13) suitable for moving in the livestock area (30), and a load sensor (14), wherein the control arrangement is configured to: - While operating the autonomous agricultural vehicle (1) according to the cleaning plan, the load amount, representing the amount of material moved by the cleaning device (13), is monitored at multiple individual times via the load sensor, and - The cleaning plan of the autonomous agricultural vehicle (1) is adjusted by adjusting the cleaning schedule of the autonomous agricultural vehicle (1) and / or one or more cleaning routes (31) based on the changes in the monitored load among the monitored loads at the multiple individual times.

18. The control arrangement of claim 17, configured to perform the method of any one of claims 2 to 16.

19. An autonomous agricultural vehicle (1) configured to clean livestock areas according to a cleaning plan, wherein, The cleaning plan defines the cleaning schedule and / or one or more cleaning routes (31) for the autonomous agricultural vehicle (1), the autonomous agricultural vehicle (1) comprising: - Motor (11), which is configured to propel the autonomous agricultural vehicle (1), - An energy storage device (12) is configured to supply energy to the motor (11). - A cleaning device (13) configured to move manure in the livestock area, - Load sensor (14), which is configured to obtain the load amount indicating the amount of material moved by the cleaning device (13), - The control arrangement according to claim 17 or 18.

20. A computer program product comprising instructions that, when a computer executes the program, cause the computer to perform the method according to any one of claims 2 to 16.

Images