Dynamic Virtual Safety Bubbles for Autonomous Cutting Vehicles

Description

1 / 87 “DYNAMIC VIRTUAL SECURITY BUBBLES FOR AUTONOMOUS CUTTING VEHICLES DESCRIPTIVE REPORT REFERENCE TO RELATED REQUESTS

[0001] This Application claims the benefit and priority of the Application United States Provisional Order No. 63 / 719,478, filed November 12, 2024, which is incorporated by reference in its entirety. BACKGROUND

[0002] This disclosure refers to the operation of a cutting vehicle in a landscaped environment, to perform one or more landscaping actions in order to maintain the landscaped environment.

[0003] A mowing vehicle typically includes one or more blades located below a cutting platform for cutting plants in a landscaped environment. The mowing vehicle typically includes a seat or cab where an operator can sit while operating the mowing vehicle. A technical challenge arises when attempting to operate a mowing vehicle autonomously, as the vehicle's blades can pose a significant risk to bystanders. Consequently, there is a need for preventative measures to minimize collisions between the vehicle and various types of objects, whether people or other objects. SUMMARY

[0004] A cutting vehicle traverses a landscaped environment and performs one or more actions in the landscaped environment. Actions may include the movement of the cutting vehicle, landscaping actions to maintain the Petition 870250101989, dated 07 / 11 / 2025, page 5 / 122 2 / 87 landscape environment or other actions to affect plants or soil in the landscape environment. The cutting vehicle can be configured to switch between autonomous operation, manual operation, or a combination thereof. Autonomous operation may refer to performing one or more actions without human intervention. Manual operation may refer to performing one or more actions with human intervention. Hybrid operation may refer to one or more actions being performed autonomously and one or more actions being performed manually. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] FIG. 1A illustrates a front perspective view of a cutting vehicle, according to one or more embodiments.

[0006] FIG. 1B illustrates a rear perspective view of the cutting vehicle, according to one or more embodiments.

[0007] FIG. 1C illustrates a plan view of the cutting vehicle, according to one or more modalities.

[0008] FIG. 2 illustrates a block diagram of the system environment for the cutting vehicle, according to one or more modalities.

[0009] FIG. 3 illustrates a cutting vehicle with one or more inherent blind spots based on the positioning of a plurality of detection mechanisms, according to one or more modalities.

[0010] FIG. 4A illustrates a first virtual safety bubble around a cutting vehicle, according to one or more embodiments.

[0011] FIG. 4B illustrates two virtual safety bubbles around a cutting vehicle, according to one or more embodiments.

[0012] FIG. 4C illustrates three virtual safety bubbles around a cutting vehicle, according to one or more modalities.

[0013] FIG. 4D illustrates the dynamic modification of a virtual safety bubble around a cutting vehicle based on the trajectory. Petition 870250101989, dated 07 / 11 / 2025, page 6 / 122 3 / 87 planned for the vehicle, according to one or more modalities.

[0014] FIG. 5 illustrates a flowchart of the method for leveraging virtual safety bubbles during the operation of the cutting vehicle, according to one or more modalities.

[0015] FIG. 6 illustrates a verification process for the cutting vehicle detection systems, according to one or more modes.

[0016] FIG. 7 illustrates a notification that the cutting vehicle is establishing the virtual safety bubble, according to one or more modalities.

[0017] FIG. 8 illustrates a notification transmitted to a client device regarding a detected obstacle, according to one or more modes.

[0018] FIG. 9 illustrates the actions that the cutting vehicle can implement when detecting an object in the virtual safety bubble, according to one or more modes.

[0019] FIG. 10 illustrates the actions that the cutting vehicle can implement when detecting an object in the virtual safety bubble, according to one or more modes.

[0020] FIG. 11 illustrates a navigation workflow of a cutting vehicle, according to one or more modes.

[0021] FIG. 12 illustrates the navigation of a cutting vehicle in a straight trajectory, according to one or more modes.

[0022] FIG. 13 illustrates the navigation of a cutting vehicle off-track in a straight trajectory, according to one or more modalities.

[0023] FIG. 14 illustrates the navigation of a cutting vehicle on-track and off-center in a straight trajectory, according to one or more modalities.

[0024] FIG. 15A illustrates the navigation of a cutting vehicle off the trajectory, but perceived as being on the trajectory, according to a Petition 870250101989, dated 07 / 11 / 2025, page 7 / 122 4 / 87 or more options.

[0025] FIG. 15B illustrates the navigation of a cutting vehicle on the trajectory, but perceived as being off the trajectory, according to one or more modes.

[0026] FIG. 16 illustrates the navigation of a cutting vehicle when entering a curve on a curved trajectory, according to one or more modes.

[0027] FIG. 17 illustrates the navigation of a cutting vehicle when exiting a curve on a curved trajectory, according to one or more modes.

[0028] FIG. 18 illustrates a process flowchart describing collision avoidance with object recognition by an autonomous cutting vehicle, according to one or more modes.

[0029] FIG. 19 illustrates an example of a decision tree for determining the size of the virtual buffer based on the characteristics of the object, according to one or more modes.

[0030] FIG. 20 illustrates a workflow for implementing virtual safety bubbles for a human detected near the cutting vehicle, according to one or more modalities.

[0031] FIG. 21 illustrates a workflow for implementing virtual safety bubbles to resolve a non-human obstacle detected near the cutting vehicle, according to one or more modalities.

[0032] FIG. 22 illustrates a workflow for human verification with multiple human detection concordance, according to one or more modalities.

[0033] FIG. 23 illustrates an example of a computer system, according to one or more modalities.

[0034] The figures represent various modalities for illustrative purposes only. An expert in the field will readily recognize, from the Petition 870250101989, dated 07 / 11 / 2025, page 8 / 122 5 / 87 discussion follows, what alternative implementation methods for the structures and approaches illustrated here can be employed without departing from the principles described here. DETAILED DESCRIPTION I. INTRODUCTION

[0035] A vehicle (e.g., a cutting vehicle) includes one or more sensors that capture information about the surroundings as the vehicle moves through an environment. The environment may include various objects (e.g., soil and obstructions) used to determine actions (e.g., a movement action to move the vehicle, perform a landscaping action to maintain a state of a landscaped environment, modify a parameter for a landscaping plan, modify an operational parameter, and modify a sensor parameter, etc.) for the vehicle to operate in the environment. A landscaping action is a physical action that alters some state of the landscaping in the environment. The landscaping action may affect plants, soil, mulch, wood chips, gravel, rocks, environment leveling, debris, or any other landscaping features (e.g., water resources, structures, etc.).

[0036] The vehicle includes a control system that processes information obtained by the sensors to generate corresponding actions. For example, the control system processes information to identify objects and generate corresponding landscaping actions. There are many examples of vehicles processing visual information obtained by an image sensor attached to the vehicle to identify environmental conditions, plan landscaping actions, identify and avoid obstructions, or a combination of both.

[0037] To assist in the safe navigation of the cutting vehicle, the control system can generate and maintain a virtual safety bubble that is Petition 870250101989, dated 07 / 11 / 2025, page 9 / 122 6 / 87 is triggered when an obstacle breaches it. When an object is determined to have breached, i.e., is within the virtual safety bubble, the control system may terminate or cease operations and / or implement other preventive measures. Preventive measures include redirecting the cutting vehicle around the obstacle, changing the cutting vehicle's configuration, requesting information from an operator or manager, etc.

[0038] The control system generates the virtual safety bubble based on the cutting vehicle's configuration. As the cutting vehicle changes configuration, the control system can automatically and / or dynamically adjust the virtual safety bubble, adjusting its characteristics. For example, when the cutting vehicle accelerates to a higher speed, the control system can automatically and / or dynamically adjust the size of the virtual safety bubble to a larger size than before, in order to provide additional distance for performing preventive measures. In another example, the cutting vehicle may change its configuration to perform different landscaping actions, and the control system can automatically and / or dynamically adjust the parameters of the virtual safety bubble in response to the changed configuration. II. ENVIRONMENTAL MANAGEMENT OF LANDSCAPE DESIGN AND LANDSCAPE PLANS Environmental management of landscape design.

[0039] Operators (“managers”) are responsible for managing landscaping operations in one or more landscape environments. Managers work to implement a landscaping objective in these environments and select from a variety of landscaping actions to implement that objective. Traditionally, managers are, for example, a landscape architect or a landscape environment manager who works with Petition 870250101989, dated 07 / 11 / 2025, page 10 / 122 7 / 87 landscaping, but they can also be other people and / or systems configured to manage landscaping operations within the landscape environment. For example, a manager could be an automated cutting vehicle, a computational machine learning model, etc. In some cases, a manager might be a combination of the managers described above. For example, a manager might include an operator assisted by a machine learning model and one or more automated cutting vehicles, or it might be an operator working in conjunction with the cutting vehicles.

[0040] Managers implement one or more landscaping objectives for a landscaped environment. A landscaping objective is typically a macro-level goal for a landscaped environment. For example, macro-level landscaping objectives might include mowing grass or plants, applying fertilizer to grass or plants, sweeping debris from the landscaped environment, or any other suitable landscaping objective. However, landscaping objectives can also be a micro-level goal for the landscaped environment. For example, micro-level landscaping objectives might include performing a small-scale action in the landscaped environment, repairing or fixing a part of a mowing vehicle, soliciting feedback from a manager, etc. Of course, there are many possible landscaping objectives and combinations of landscaping objectives, and the examples described above are not intended to be limiting.

[0041] In one or more modalities, landscaping objectives are achieved by one or more cutting vehicles performing a series of landscaping actions. Cutting vehicles are described in greater detail below. Landscaping actions can be any operation implementable by a cutting vehicle within the landscape environment that fulfills a landscaping objective. Consider, for example, a landscaping objective of maintaining a golf course. This landscaping objective requires Petition 870250101989, dated 07 / 11 / 2025, page 11 / 122 8 / 87 a series of landscaping actions, for example, mowing different parts of the field to different grass heights, fertilizing parts of the field, repairing parts of the field, etc. Similarly, each landscaping action belonging to the overall objective can be a landscaping objective in itself. For example, mowing the lawn may require its own set of landscaping actions, for example, clearing debris, mowing the lawn, applying new seeds, etc.

[0042] In other words, managers implement a landscaping plan in the landscape environment to achieve a landscaping objective. A landscaping plan is a hierarchical set of macro and / or micro-level objectives that achieve the manager's landscaping objective. In a landscaping plan, each macro or micro-objective may require a set of landscaping actions to be achieved, or each macro or micro-objective may itself be a landscaping action. Thus, to expand, the landscaping plan is a temporally sequenced set of landscaping actions to be applied to the landscape environment that the manager expects to achieve the landscaping objective.

[0043] When executing a landscaping plan in a landscaped environment, the landscaping plan itself and / or its constituent landscaping objectives and actions produce various outcomes. An outcome is a representation of whether, or how well, a mowing vehicle achieved the landscaping plan, the landscaping objective, and / or the landscaping action. An outcome can be a qualitative measure, such as “achieved” or “not achieved,” or it can be a quantitative measure, such as “1.5 acres (6070.28 m2) mowed.” Outcomes can also be positive or negative, depending on the mowing vehicle configuration or the landscaping plan implementation. Furthermore, outcomes can be measured by sensors on the mowing vehicle, entered by managers, or accessed via a database or network.

[0044] Traditionally, managers have leveraged their experience, Petition 870250101989, dated 07 / 11 / 2025, page 12 / 122 9 / 87 expertise and technical knowledge when implementing landscaping actions in a landscaping plan. For example, a manager may rely on established best practices to determine a specific set of landscaping actions to be performed in a landscaping plan to achieve a landscaping objective. Other examples include leveraging their expertise in the order of actions or workflow.

[0045] Leveraging historical and managerial knowledge to make decisions about a landscape plan affects the spatial and temporal characteristics of a landscape plan. For example, landscaping actions in a landscape plan have historically been applied to the entire landscape environment, rather than small parts of it. For instance, in the grand scheme of landscape environment maintenance, the manager might plan the entire schedule of actions that must be executed to achieve the goal of maintaining a landscape environment. Similarly, each landscaping action in the sequence of landscaping actions of a landscape plan is historically performed at approximately the same time. For example, when a manager decides to fertilize a landscape environment, they fertilize the landscape environment at approximately the same time; or, when the manager decides to mow the lawn of the landscape environment, they do so at approximately the same time.

[0046] Notably, however, mowing vehicles have advanced significantly in their capabilities. For example, mowing vehicles continue to become more autonomous, include an increasing number of sensors and measuring devices, employ greater amounts of processing power and connectivity, and implement various computer vision algorithms to enable managers to successfully implement a landscaping plan.

[0047] Due to this increased capacity, managers are no longer limited to spatially and temporally monolithic implementations of Petition 870250101989, dated 07 / 11 / 2025, page 13 / 122 10 / 87 landscaping actions in a landscaping plan. Instead, managers can leverage the advanced capabilities of cutting vehicles to implement highly localized landscaping plans determined by real-time measurements in the landscape environment. In other words, instead of applying a "best estimate" landscaping plan to an entire landscape environment, they can implement individualized and informed landscaping plans within the landscape environment. III. CUTTING VEHICLE Overview

[0048] A cutting vehicle that implements landscaping actions from a landscaping plan can have a variety of configurations, some of which are described in greater detail below.

[0049] The cutting vehicle generally includes a detection mechanism, a landscaping mechanism, and a control system. The cutting vehicle may additionally include a wireless transmitter, a display, manual controls, a power source, digital memory, a communication device, or any other suitable component that allows the cutting vehicle to implement landscaping actions in a landscaping plan. Furthermore, the components and functions of the cutting vehicle described are only examples, and a cutting vehicle may have different or additional components and functions beyond those described below. Operating environment

[0050] The cutting vehicle operates in an operational environment. The operational environment is the environment around the cutting vehicle as it implements landscaping actions of a landscaping plan. The operational environment may also include the cutting vehicle itself and its Petition 870250101989, dated 07 / 11 / 2025, page 14 / 122 11 / 87 matching components.

[0051] The operational environment typically includes a landscaped environment, and the cutting vehicle usually implements landscaping actions from the landscaping plan in the landscaped environment. A landscaped environment is a geographical area where the cutting vehicle implements a landscaping plan. The landscaped environment can be an external landscaped environment, but it can also be an indoor location or any other suitable environment. The landscaped environment may include a lawn, other plant landscaping features, stones, gravel, mulch, wood chips, water features, etc.

[0052] A landscape environment can include any number of landscape environment parts. A landscape environment part is a subunit of a landscape environment. For example, a landscape environment part might be a landscape environment part designated for a landscape structure. Or, in another example, a landscape environment part might be driveways for vehicle movement. The cutting vehicle 100 can perform different landscaping actions for different parts of the landscape environment. For example, the cutting vehicle 100 might cut grass in one part of the landscape environment while removing debris in another part of the landscape environment. Furthermore, a landscape environment and a landscape environment part are largely interchangeable in the context of the methods and systems described herein.In other words, landscape plans and their corresponding landscaping actions can be applied to an entire landscape environment or to a part of the landscape environment, depending on the circumstances in question. III.A EXEMPLARY CONFIGURATIONS Detection mechanism(s) Petition 870250101989, dated 07 / 11 / 2025, p. 15 / 122 12 / 87

[0053] The cutting vehicle may include a detection mechanism. The detection mechanism identifies objects in the cutting vehicle's operating environment. To do this, the detection mechanism obtains information describing the environment (e.g., sensor or image data) and processes this information to identify relevant objects (e.g., obstacles, people, other vehicles, etc.) in the operating environment. Identifying objects in the environment further allows the cutting vehicle to implement landscaping actions in the landscaped environment. For example, the detection mechanism may capture an image of the landscaped environment and process it to identify any human operators in the vehicle's vicinity. The cutting vehicle then implements landscaping actions in the landscaped environment based on the identified objects, for example, avoiding collisions with obstacles in the environment.

[0054] The cutting vehicle may include any number or type of sensing mechanism that may assist in determining and implementing landscaping actions. In some embodiments, the sensing mechanism includes one or more sensors. For example, the sensing mechanism may include a multispectral camera, a stereo camera, a CCD camera, a single-lens camera, a CMOS camera, a hyperspectral imaging system, a LIDAR (light sensing and ranging system), a depth sensing system, a dynamometer, an infrared camera, a thermal camera, a humidity sensor, a light sensor, a temperature sensor, an inertial measurement unit (IMU) sensor, an accelerometer, a sensor coupled to one or more sets of motors that control the movement of the vehicle or its components, or any other suitable sensor.In addition, the detection mechanism may include a set of sensors (e.g., a set of cameras) configured to capture information about the environment surrounding the cutting vehicle 100. For example, the mechanism of... Petition 870250101989, dated 07 / 11 / 2025, page 16 / 122 13 / 87 detection may include a set of cameras configured to capture a set of images representing the environment around the mowing vehicle. The detection mechanism may also be a sensor that measures the state of the mowing vehicle. For example, the detection mechanism may be a speed sensor, a temperature sensor, a wheel sensor, a fuel level sensor, a battery level sensor, or some other sensor that can monitor the state of a component of the mowing vehicle. Additionally, the detection mechanism may also be a sensor that measures components during the implementation of a landscaping action. Whatever the case, the detection mechanism detects information about the operating environment (including the mowing vehicle).

[0055] A detection mechanism can be mounted at any point on the mounting mechanism. Depending on where the detection mechanism is mounted relative to the landscaping mechanism, one or the other may pass over a geographic area in the landscape environment before the other. For example, the detection mechanism may be positioned on the mounting mechanism so that it passes through a geographic location before the landscaping mechanism as the cutting vehicle moves through the landscape environment. In other examples, the detection mechanism is positioned on the mounting mechanism so that both pass through a geographic location at substantially the same time as the cutting vehicle moves through the field. Similarly, the detection mechanism may be positioned on the mounting mechanism so that the landscaping mechanism passes through a geographic location before the detection mechanism as the cutting vehicle moves through the landscape environment.The detection mechanism can be statically mounted on the mounting mechanism or it can be removablely or dynamically coupled to the mounting mechanism. In other examples, the detection mechanism can be mounted on some other surface of the... Petition 870250101989, dated 07 / 11 / 2025, page 17 / 122 14 / 87 cutting vehicle or can be incorporated into another component of the cutting vehicle. The detection mechanism can be removablely attached to the cutting vehicle. Verification mechanism(s)

[0056] The cutting vehicle may include a verification mechanism. Generally, the verification mechanism records a measurement of the operating environment, and the cutting vehicle can use the recorded measurement to verify or determine the extent of an implemented landscaping action (i.e., an outcome of the landscaping action).

[0057] To illustrate, consider an example where a cutting vehicle implements a landscaping action based on a measurement of the operating environment by the detection mechanism. The verification mechanism records a measurement of the same geographic area measured by the detection mechanism and where the cutting vehicle implemented the determined landscaping action. The cutting vehicle then processes the recorded measurement to determine the outcome of the landscaping action. For example, the verification mechanism might record an image of the geographic region around a leveled portion of terrain identified by the detection mechanism and processed by a landscaping mechanism. The cutting vehicle might apply a detection algorithm to the recorded image to determine the outcome of the landscaping action.

[0058] The information recorded by the verification mechanism can also be used to empirically determine the operating parameters of the cutting vehicle that will achieve the desired effects of the implemented landscaping actions (e.g., to calibrate the cutting vehicle, to modify landscaping plans, etc.). For example, the cutting vehicle can apply a calibration detection algorithm to a measurement recorded by the cutting vehicle. In this case, the cutting vehicle determines whether Petition 870250101989, dated 07 / 11 / 2025, page 18 / 122 15 / 87 The actual effects of an implemented landscaping action are the same as its intended effects. If the effects of the implemented landscaping action differ from its intended effects, the cutting vehicle can perform a calibration process. The calibration process alters the operating parameters of the cutting vehicle so that the effects of future implemented landscaping actions are the same as their intended effects. To illustrate, consider the previous example, where the cutting vehicle recorded an image of an object in the landscaping environment. In this case, the cutting vehicle can apply a calibration algorithm to the recorded image to determine if the construction is properly calibrated (e.g., in its intended location in the operating environment).If the cutting vehicle determines that it is not calibrated (for example, landscaping work has resulted in some imprecision or inaccuracy), the cutting vehicle can self-calibrate so that the futures are in the correct location. Other examples of calibrations are also possible.

[0059] The verification mechanism can have various configurations. For example, the verification mechanism can be substantially similar (e.g., be of the same type of mechanism) to the detection mechanism, or it can be different from the detection mechanism. In some cases, the detection mechanism and the verification mechanism can be the same (e.g., the same sensor). In an exemplary configuration, the verification mechanism is positioned distally to the detection mechanism relative to the direction of travel, and the landscaping mechanism is positioned between them. In this configuration, the verification mechanism traverses a geographical location in the operating environment after the landscaping mechanism and the detection mechanism. However, the mounting mechanism can maintain the relative positions of the system components in any other suitable configuration. In Petition 870250101989, dated 07 / 11 / 2025, p. 19 / 122 16 / 87 In some configurations, the verification mechanism may be included in other components of the cutting vehicle.

[0060] The cutting vehicle may include any number or type of verification mechanism. In some embodiments, the verification mechanism includes one or more sensors. For example, the verification mechanism may include a multispectral camera, a stereo camera, a CCD camera, a single-lens camera, a CMOS camera, a hyperspectral imaging system, a LIDAR (light detection and ranging system), a depth sensing system, a dynamometer, an infrared camera, a thermal camera, a humidity sensor, a light sensor, a temperature sensor, or any other suitable sensor. In addition, the verification mechanism may include a sensor array (e.g., a camera array) configured to capture information about the environment surrounding the cutting vehicle. For example, the verification mechanism may include a camera array configured to capture an image array representing the operating environment. Landscaping mechanism(s)

[0061] The cutting vehicle may include a landscaping mechanism. The landscaping mechanism can implement landscaping actions in the operational environment of a cutting vehicle. For example, a cutting vehicle may include a landscaping mechanism that performs one or more useful physical actions to achieve landscaping objectives, i.e., landscaping actions. In some embodiments, the landscaping mechanism may be configured to perform a plurality of different landscaping actions. In other embodiments, the landscaping mechanism may be specific to perform one type of landscaping action.

[0062] Depending on the configuration, the cutting vehicle may include Petition 870250101989, dated 07 / 11 / 2025, page 20 / 122 17 / 87 various numbers of landscaping mechanisms (e.g., 1, 2, 5, 20, etc.). A landscaping mechanism may be fixed (e.g., statically coupled) to the mounting mechanism or coupled to the cutting vehicle. Alternatively, or additionally, a landscaping mechanism may be mobile (e.g., translatable, rotatable, etc.) on the cutting vehicle. In one configuration, the cutting vehicle includes a single landscaping mechanism. In this case, the landscaping mechanism may be actuatable to align it to a specific position and / or orientation. In a second variation, the cutting vehicle includes an array of landscaping mechanisms comprising an array of landscaping mechanisms. In this configuration, a landscaping mechanism may be a single landscaping mechanism, a combination of landscaping mechanisms, or the entire array of landscaping mechanisms.Thus, a single landscaping mechanism, a combination of landscaping mechanisms, or the entire set can be selected to perform landscaping actions. Similarly, the single set, the combination, or the entire set can be triggered to align with a landscaped environment as needed. In some configurations, the cutting vehicle can align a landscaping mechanism to an object identified in the operational environment. That is, the cutting vehicle can identify an object in the operational environment and trigger the landscaping mechanism so that its landscaped environment aligns with the identified object.

[0063] A landscaping mechanism can be operable between different modes. For example, in one mode, the landscaping mechanism may be on standby and not performing any action at the moment. In another mode, the landscaping mechanism may be performing an action. In yet another mode, the landscaping mechanism may be performing a different type of action.

[0064] The configuration of the landscaping mechanism can affect the Petition 870250101989, dated 07 / 11 / 2025, page 21 / 122 18 / 87 parameters of the virtual safety bubble. For example, the landscaping mechanism can be retracted into a compact configuration or deployed in an expanded configuration, and the control system that generates the virtual safety bubble can automatically and / or dynamically adjust the parameters of the safety bubble based on whether the landscaping mechanism is in the retracted or expanded configuration. In another example, the landscaping mechanism can be operable for multiple landscaping actions. Based on the landscaping action, the control system can automatically and / or dynamically adjust the parameters of the safety bubble.

[0065] In one or more embodiments, the landscaping mechanism may be a device for cutting grass or other plants to a target height. The landscaping mechanism is capable of cutting grass or other plants to a range of heights. An operator can set the height at which to cut the plant. The landscaping mechanism may be a cutting platform with one or more rotating blades positioned below the vehicle. As the rotating blades turn, they capture and cut the plant. The vertical height movement of the rotating blade, i.e., relative to the vehicle chassis, affects the cutting height. Control system(s)

[0066] The mowing vehicle includes a control system. The control system controls the operation of the various components and systems of the mowing vehicle. For example, the control system may obtain information about the operating environment, process this information to identify a landscaping action to be implemented, and implement the identified landscaping action with components of the mowing vehicle's system. The control system may further assist in navigating the mowing vehicle through the operating environment. Navigation may include data collection and analysis. Petition 870250101989, dated 07 / 11 / 2025, page 22 / 122 19 / 87 related to the environment from one or more sensors and the generation of navigation instructions based on the data. More details about the control system are included below in other figures.

[0067] The control system can receive information from the detection mechanism, the verification mechanism, the landscaping mechanism, and / or any other component or system of the cutting vehicle. For example, the control system can receive measurements from the detection mechanism or the verification mechanism, or information related to the state of a landscaping mechanism or landscaping actions implemented from a verification mechanism. Other information is also possible.

[0068] Similarly, the control system can provide input data to the detection mechanism, the verification mechanism, and / or the landscaping mechanism. For example, the control system can be configured to input and control operational parameters of the cutting vehicle (e.g., speed, direction). Similarly, the control system can be configured to input and control operational parameters of the detection mechanism and / or verification mechanism. The operational parameters of the detection mechanism and / or verification mechanism may include processing time, location and / or angle of the detection mechanism, image capture intervals, image capture settings, etc. Other inputs are also possible. Finally, the control system can be configured to generate machine input data for the landscaping mechanism.This means translating a landscaping action from a landscaping plan into machine instructions that can be implemented by the landscaping mechanism.

[0069] The control system can be operated by a user operating the cutting vehicle, in a fully or partially autonomous manner, operated by a user connected to the cutting vehicle via a network, or any Petition 870250101989, dated 07 / 11 / 2025, page 23 / 122 20 / 87 combination of the above items. For example, the control system can be operated by an operator seated in the cab of the cutting vehicle, or it can be operated by an operator connected to the control system via a wireless network. In another example, the control system can implement a series of control algorithms, computer vision algorithms, decision algorithms, etc., that allow it to operate autonomously or partially autonomously.

[0070] The control system can be implemented by a computer or a distributed system of computers. The computers can be connected in various network environments. For example, the control system could be a series of computers implemented in the cutting vehicle and connected by a local area network. In another example, the control system could be a series of computers implemented in the cutting vehicle, in the cloud, on a client device, and connected by a wireless network.

[0071] The control system can apply one or more computational models to determine and implement landscaping actions in the landscape environment. For example, the control system can apply an object detection model to images acquired by the detection mechanism to identify and classify objects in the sensor data. Based on the detected objects, the control system can determine parameters for the landscaping actions to be performed by the landscaping mechanism. The control system can be coupled to the cutting vehicle so that an operator (e.g., a driver) can interact with the control system. In other embodiments, the control system is physically removed from the cutting vehicle and communicates with the system components (e.g., detection mechanism, landscaping mechanism, etc.) wirelessly.

[0072] In some configurations, the cutting vehicle may include Petition 870250101989, dated 07 / 11 / 2025, page 24 / 122 21 / 87 additionally a communication device, which functions to communicate (e.g., send and / or receive) data between the control system and a set of remote devices. The communication device may be a Wi-Fi communication system, a cellular communication system, a short-range communication system (e.g., Bluetooth, NFC, etc.) or any other suitable communication system. Other Machine Components

[0073] In various configurations, the cutting vehicle may include any number of additional components.

[0074] For example, the cutting vehicle may include a mounting mechanism. The mounting mechanism provides a mounting point for the cutting vehicle components. That is, the mounting mechanism may be a chassis or frame to which the cutting vehicle components can be attached, but it may alternatively be any other suitable mounting mechanism. More generally, the mounting mechanism statically retains and mechanically supports the positions of the detection mechanism, the landscaping mechanism, and the checking mechanism. In one exemplary configuration, the mounting mechanism extends outward from the cutting vehicle body so that the mounting mechanism is approximately perpendicular to the direction of travel. In some configurations, the mounting mechanism may include a set of landscaping mechanisms positioned laterally along the mounting mechanism.In some configurations, the cutting vehicle may not include a mounting mechanism, the mounting mechanism may be positioned alternatively, or the mounting mechanism may be incorporated into any other component of the cutting vehicle.

[0075] The cutting vehicle may include propulsion mechanisms. The Petition 870250101989, dated 07 / 11 / 2025, page 25 / 122 22 / 87 Locomotion mechanisms may include any number of wheels, continuous tracks, articulated legs, or some other locomotion mechanism. For example, the cutting vehicle may include a first set and a second set of coaxial wheels, or a first set and a second set of continuous tracks. In both examples, the axes of rotation of the first and second sets of wheels / tracks are approximately parallel. Furthermore, each set is arranged along opposite sides of the cutting vehicle. Typically, the locomotion mechanisms are coupled to a drive mechanism that causes the locomotion mechanisms to translate the cutting vehicle through the operating environment. For example, the cutting vehicle may include a drive assembly to rotate wheels or tracks. In different configurations, the cutting vehicle may include any other suitable number or combination of locomotion mechanisms and drive mechanisms.

[0076] The cutting vehicle may also include one or more coupling mechanisms (e.g., a hitch). The coupling mechanism functions to couple, either removablely or statically, various components of the cutting vehicle. For example, a coupling mechanism may couple a drive mechanism to a secondary component, so that the secondary component is pulled to the rear of the cutting vehicle. In another example, a coupling mechanism may couple one or more landscaping mechanisms to the cutting vehicle.

[0077] The cutting vehicle may additionally include a power source, which functions to power the system components, including the detection mechanism, the control system, and the landscaping mechanism. The power source may be mounted on the mounting mechanism, or may be removablely coupled to the mechanism. Petition 870250101989, dated 07 / 11 / 2025, page 26 / 122 23 / 87 mounting or may be incorporated into another component of the system (e.g., located in the drive mechanism). The power source may be a rechargeable power source (e.g., a set of rechargeable batteries), a power source for energy harvesting (e.g., a solar system), a fuel-consuming power source (e.g., a fuel cell assembly or an internal combustion system), or any other suitable power source. In other configurations, the power source may be incorporated into any other component of the cutting vehicle. III.B EXEMPLARY CUTTING VEHICLE

[0078] FIGS. 1A-1C illustrate various views of a cutting vehicle. 100 is an example, according to one or more modalities. The cutting vehicle 100 is configured to perform a cutting action. The cutting vehicle 100 can operate autonomously, manually, or in a hybrid manner.

[0079] FIG. 1A illustrates a front perspective view of the cutting vehicle 100, according to one or more embodiments. The front view illustrates the cutting vehicle 100, including the detection mechanisms 110A, 110B, 110C and 110D, a cutting platform 120, a control system 130 and a transmitter 140.

[0080] Detection mechanisms 110A, 110B, 110C, and 110D can measure or detect information describing the operating environment of the cutting vehicle 100. In one or more embodiments, the detection mechanisms 110 can be cameras, LIDAR sensors, other depth and range sensors, or a combination thereof. In the embodiment shown, there are two detection mechanisms 110A and 110B (e.g., cameras) oriented forward, along a primary direction of travel. Two detection mechanisms 110C and 110D (e.g., cameras) are oriented to the left side of the vehicle. Petition 870250101989, dated 07 / 11 / 2025, p. 27 / 122 24 / 87

[0081] The cutting platform 120 is an embodiment of a landscaping mechanism. In the embodiment shown, the cutting platform 120 is situated near the underside of the cutting vehicle 100, close to the ground. The cutting platform 120 may include a large guard, under which are one or more motorized blades for cutting plants in the environment. Each of the motorized blades may be individually addressable and operable by the cutting vehicle 100. The cutting platform 120 may be coupled to a chute configured to guide the debris ingested by the cutting vehicle 100, i.e., the cutting platform 120, in a specific direction. In some embodiments, a container may be implemented and coupled to the chute, such that the container is configured to store debris from the cutting vehicle 100 during operation. The motorized blades may also include a braking system to stop the rotation of the blades.

[0082] The control system 130 analyzes the data received by the detection mechanisms 110 and generates control signals to control the operation of the cutting vehicle 100. In one or more embodiments, the control system is further configured to generate one or more virtual safety bubbles for safe navigation of the cutting vehicle 100. The control system 130 can generate one or more virtual safety bubbles based on the configuration of the cutting vehicle 100. The generation of the safety bubbles is described in more detail in FIGS. 3 to 10. With the safety bubbles, the control system 130 can identify and track objects in the operating environment to determine if the objects in the operating environment would create a safety problem, for example, typically if the cutting vehicle 100 and the object are designed to collide.If the object interacts with the virtual safety bubble, the control system 130 can implement responsive measures to promote safe navigation. Examples of responsive measures are described below in the FIGS. Petition 870250101989, dated 07 / 11 / 2025, page 28 / 122 25 / 87 from 11 to 19. The control system 130 can be positioned anywhere on the cutting vehicle 100, for example, internally or externally, etc.

[0083] Transmitter 140 enables wireless communication with other devices or computing systems, for example, via a wireless network. Transmitter 140 can be a component of control system 130. Transmitter 140 can be configured to receive and transmit communications.

[0084] In some embodiments, the cutting vehicle 100 also includes an alert system to generate alerts to warn any passersby in the operating environment. The alert system may include a loudspeaker system to generate audible alerts, one or more light emitters to generate visual alerts, another type of emitter, or a combination thereof.

[0085] FIG. 1B illustrates a rear perspective view of the cutting vehicle 100, according to one or more embodiments. The cutting vehicle 100 may include additional detection mechanisms 110E, 110F, 110G and 110H. Detection mechanisms 110E and 110F are oriented to the rear. Detection mechanisms 110G and 110H are oriented to the right of the vehicle. The cutting vehicle 100 may also include a platform 150 so that an operator can accompany and / or operate it manually.

[0086] FIG. 1C illustrates a top-down view of the cutting vehicle 100, according to one or more embodiments. The top-down view further illustrates the manual controls 170 and a display 180.

[0087] The manual controls 170 may include a switch to toggle the cutting vehicle 100 from autonomous operation to manual operation, or vice versa. The manual controls 170 may also include other manual inputs to control the operation of the cutting vehicle 100, for example, to navigate the cutting vehicle 100, to activate the cutting platform 120, to activate the warning system, etc.

[0088] The 180 display is an electronic display configured to provide Petition 870250101989, dated 07 / 11 / 2025, page 29 / 122 26 / 87 visual content to the operator. Display 180 can show the status of cutting vehicle 100. Display 180 can also have a user interface, which can present interactive elements so that the operator can provide additional manual inputs in the operation of cutting vehicle 100. III.C SYSTEM ENVIRONMENT

[0089] FIG. 2 is a block diagram of the system environment 200 for the cutting vehicle 100, according to one or more embodiments. In this example, the control system 210 (e.g., control system) is connected to external systems 220 and a set of machine components 230 via a network 240 within the system environment 200.

[0090] External systems 220 are any systems that can generate data representing useful information for determining and implementing landscaping actions in a landscape environment. External systems 220 may include one or more sensors 222, one or more processing units 224, and one or more data repositories 226. The sensor(s) 222 may measure the landscape environment, the operational environment, the cutting vehicle 100, etc., and generate data representing these measurements. For example, sensors 222 might include a precipitation sensor, a wind sensor, a heat sensor, a camera, etc.The 2240 processing units can process measured data to provide additional information that can assist in determining and implementing landscaping actions in the landscape environment. The 226 data repositories store historical information about the cutting vehicle 100, the operating environment, the landscape environment, etc., which can be useful in determining and implementing landscaping actions in the landscape environment. For example, the 226 data repository can store results of previously implemented landscaping plans and landscaping actions for a landscape environment. Petition 870250101989, dated 07 / 11 / 2025, page 30 / 122 27 / 87 nearby landscape and / or region. Historical information may have been obtained from one or more cutting vehicles (i.e., measuring the result of a landscaping action of a first cutting vehicle with the sensors of a second cutting vehicle). Additionally, data storage 226 may store results of specific landscaping actions in the landscape environment or results of landscaping actions performed in nearby landscape environments with similar characteristics. Data storage 226 may also store historical data on climate, flooding, landscape environment use, completed operations, scheduled operations, etc., for the landscape environment and the surrounding area. Finally, data storage 226 may store any information measured by other components in the system environment 200. External systems 220 may be implemented in the cloud.

[0091] The machine component set 230 includes one or more components 232. Components 222 are elements of the cutting vehicle 100 that can perform landscaping actions (e.g., a landscaping mechanism). As illustrated, each component has one or more input controllers 234 and one or more sensors 236, but a component may include only sensors 236 or only input controllers 234. An input controller 234 controls the function of component 232. For example, an input controller 234 may receive commands from the machine via the network 240 and trigger component 230 in response. A sensor 226 generates data representing measurements of the operating environment and provides this data to other systems and components within the system environment 200. The measurements may be of a component 232, the cutting vehicle 100, the operating environment, etc.For example, a sensor 226 can measure a configuration or state of component 222 (e.g., a configuration, parameter, load of). Petition 870250101989, dated 07 / 11 / 2025, page 31 / 122 28 / 87 energy etc.), measure conditions in the operating environment (e.g., humidity, temperature, etc.), capture information representing the operating environment (e.g., images, depth information, distance information), and generate data representing the measurement(s).

[0092] The control system 210 receives information from external systems 220 and the machine component set 230 and implements a landscaping plan in a landscaped environment using a cutting vehicle 100. Before implementing the landscaping plan, the cutting vehicle checks if it is safe to operate. To do this, the control system 210 receives a notification from a manager stating that the environment around the cutting vehicle is safe to operate and free of obstacles. The control system 210 checks, using captured images, if there are no obstacles in the environment around the cutting vehicle.The 210 control system generates a virtual safety bubble for landscaping operations based on the mowing vehicle's configuration. While the mowing vehicle performs the landscaping operations, the 210 control system continuously identifies and locates obstacles in the environment. If an obstacle falls within the virtual safety bubble, the 210 control system can either stop the operation or implement preventative measures.

[0093] The control system 210 includes a safety bubble generation module 212, a classification module 214, a safety module 216, a navigation module 218, and a user interface module 219. In other embodiments, the control system 210 has additional / fewer modules. In other embodiments, the modules can be variably configured so that the functions of one module can be performed by one or more other modules.

[0094] The safety bubble generation module 212 generates a virtual safety bubble for the cutting vehicle 100. The virtual safety bubble can have a three-dimensional shape around the cutting vehicle 100. Petition 870250101989, dated 07 / 11 / 2025, page 32 / 122 29 / 87 In other configurations, the virtual safety bubble may take the form of a strip, for example, a wall of a certain height surrounding the cutting vehicle 100. Various other shapes and sizes can be considered. The safety bubble generation module 212 defines the shape and size of the virtual safety bubble based on the configuration of the cutting vehicle 100. For example, the safety bubble generation module 212 can determine the shape and / or size of the virtual safety bubble based on whether the cutting vehicle 100 is in a first configuration to navigate to an operational environment or in a second configuration to execute a landscaping plan. The safety bubble generation module 212 can dynamically adjust the virtual safety bubble based on sensor data.For example, the 212 security bubble generation module can increase the size of the virtual security bubble in response to the darkening of the operating environment due to sunset.

[0095] The classification module 214 classifies objects in images captured by the cameras (embedding sensors 222) implemented in the cutting vehicle 100. The classification module 214 can use one or more models to classify pixels related to objects in the image. One model can identify obstacles as objects that are not part of the landscaping operation. For example, the model might classify rows of crops as non-obstacles, but would classify a wild fox or a large rock as an obstacle. Another model can perform image segmentation, classifying pixels for various types of objects, for example, the ground, the sky, foliage, obstacles, etc. Another model can calculate the speed of objects relative to the cutting vehicle 100, for example, using one or more visual odometry methods.And yet another model can predict the depth of objects from the camera, for example, using a depth estimation model trained for it. Petition 870250101989, dated 07 / 11 / 2025, page 33 / 122 30 / 87 predict depth based on image data. Depth generally refers to the distance between the clipping vehicle and pixels or objects in the images. For example, a first object present in an image can be determined to be at a depth of 5 meters from the clipping vehicle. The 214 classification module can further generate 3D point cloud representations of objects within a virtual operating environment, allowing object tracking. The various models can incorporate other sensor data (captured by sensors 222 or 236) to aid in classification, for example, LIDAR data, temperature measurements, etc.

[0096] Safety module 216 assesses whether obstacles are within the virtual safety bubble. Safety module 216 can use a depth estimation model to predict the depths of obstacles relative to cutting vehicle 100. If an obstacle has a depth below the virtual safety, i.e., if any part of the obstacle breaks through the virtual safety bubble barrier, safety module 216 sends this warning to navigation module 218, for example, to stop operation or implement preventive measures.

[0097] Navigation module 218 navigates through cutting vehicle 100. Navigation module 218 generates navigation instructions based on a landscaping plan. The landscaping plan may include one or more landscaping operations to be completed. Navigation module 218 can plot a route for the vehicle to navigate. Navigation module 218 can adjust the navigation route based on sensor data. Navigation module 218 can receive notifications from safety module 216 that an obstacle has crossed the virtual safety bubble. In response to the notification, navigation module 218 can interrupt operations, implement other preventive measures, or a combination thereof. As an example of a preventive measure, navigation module 218 Petition 870250101989, dated 07 / 11 / 2025, page 34 / 122 31 / 87 can stop the cutting vehicle 100 when a warning is issued that an obstacle has passed the virtual safety bubble. As another example of a preventive measure, navigation module 218 can plot a route around the obstacle to avoid collisions. Additional details related to navigation by navigation module 218 are described in FIGS. 11 to 19.

[0098] User interface module 219 maintains a graphical user interface (GUI) to display information to the cutting vehicle 100 manager and receive input from the manager. User interface module 219 can graphically illustrate the cutting vehicle 100 in operation, for example, when moving along a path or when performing one or more landscaping actions. The GUI can also display any obstacles or other objects in the operating environment. The GUI can further be configured to receive input to control the cutting vehicle 100. Example inputs include switching the speed of the cutting vehicle 100, manually adjusting the virtual safety bubble, etc. In one embodiment, the GUI can notify a cutting vehicle 100 manager that an obstacle has breached the virtual safety bubble. The GUI can prompt the manager for action or input on how to respond. Example user interfaces are described in more detail in FIGS. 6 to 10.

[0099] In one or more modes, the models used by the control system 110 can be trained as machine learning models using training data. Training can be supervised, unsupervised, or semi-supervised. Several types of machine learning model architectures can be implemented, for example, neural networks, decision trees, support vector machine learning, etc.

[0100] The 240 network connects nodes in the 200 system environment to allow microcontrollers and devices to communicate with each other. In Petition 870250101989, dated 07 / 11 / 2025, page 35 / 122 32 / 87 In some modes, the components are connected within the network as a Controller Area Network (CAN). In this case, within the network, each element has an input and output connection, and the network 250 can translate information between the various elements. For example, the network 250 receives input information from the camera array 210 and the component array 220, processes the information, and transmits it to the control system 230. The control system 230 generates a landscaping action based on the information and transmits instructions to implement the landscaping action to the appropriate component(s) 222 of the component array 220.

[0101] In addition, the 200 system environment may be other types of network environments and include other networks, or a combination of network environments with multiple networks. For example, the 200 system environment may be a network such as the Internet, a LAN, a MAN, a WAN, a wired or wireless mobile network, a private network, a virtual private network, a direct communication line, and the like. IV. Obstructed Views and Unobstructed Views

[0102] As described above, a cutting vehicle is configured with one or more detection mechanisms (“detection system”) to measure the environment. In one configuration, the detection system may be a set of detection mechanisms configured to capture images of the environment. The image data in the image represents the various objects in the environment surrounding the cutting vehicle. Thus, the detection system is configured to capture image data of the environment. In other examples, the detection mechanisms may capture depth data, for example, from one or more depth and range sensors.

[0103] The detection system has a field of view and, as the Petition 870250101989, dated 07 / 11 / 2025, page 36 / 122 The 33 / 87 detection system is a set of detection mechanisms; the field of view of the detection system can encompass various landscape viewing environments that can be combined to form a 360-degree view. That is, each detection mechanism has its own field of view, and the landscape viewing environments, together, form the field of view of the detection system.

[0104] There may be one or more blind spots in a field of view, caused by the configuration of the detection system. Some blind spots may include areas outside the reach of any detection mechanism and obstructed views, for example, views within the detection system's field of view but obstructed by one or more objects. Obstructed views comprise image data in images where an object obstructs an object or objects behind it (such that the obstructed objects are obscured from view). Unobstructed views comprise image data in images where no object obstructs an object or objects behind it. For example, consider a detection mechanism capturing images of a tire attached to the cutting vehicle and the surrounding environment. Because the tire is obscuring image data of objects behind it (e.g., ground, rocks, etc.), this is an obstructed view. The rest of the image is an unobstructed view because there are no objects obscuring other objects.Or, in another example, an operator traveling alongside the cutting vehicle may obstruct one or more detection mechanisms (for example, the rear detection mechanism 110C and 110D in FIG. 1B).

[0105] Obstructed views are problematic in autonomous construction due to their inherent safety issues. For example, an object that may be a significant obstacle can be obscured by another object in an obstructed view. The cutting vehicle may therefore be unable to identify and take into account a problematic obstacle. Methods are presented here for establishing a virtual safety bubble. Petition 870250101989, dated 07 / 11 / 2025, page 37 / 122 34 / 87 to prevent obstacles from obstructing the cutting vehicle's view.

[0106] FIG. 3 illustrates a cutting vehicle 300 (an embodiment of the cutting vehicle 100) equipped with a detection system. The cutting vehicle 300 has a detection system with a total of eight detection mechanisms 310. A first pair 310A of detection mechanisms is positioned near the front end of the cutting vehicle 300 and oriented forward with a field of view 315A. A second pair 310B of detection mechanisms is positioned near the rear end of the cutting vehicle 300 and oriented backward with a field of view 315B. A third pair 310C of detection mechanisms is positioned near the left side of the cutting vehicle 300 and oriented to the left with a field of view 315C. And a fourth pair 310D of detection mechanisms is positioned near the right side of the cutting vehicle 300 and oriented to the right with a field of view 315D.As noted above, the field of view of the detection system can aggregate the individual fields of view of the 315 detection mechanisms.

[0107] FIG. 3 also illustrates an obstacle 320. Obstacle 320 is located ahead, to the left, of cutting vehicle 300. Cutting vehicle 300 is configured to apply an obstacle detection model to the images captured by the detection mechanism to identify obstacles in the environment. That is, cutting vehicle 300 uses the obstacle detection model to determine which pixels in the images represent obstacles and pinpoints the approximate location of these obstacles in the environment (e.g., by estimating depth).

[0108] As described above, the 300 cutting vehicle detection system includes several blind spots. Blind spots are areas in the environment not visible to the cutting vehicle because, for example, a part of the vehicle obstructs the view (e.g., behind a tire) or the mechanisms of Petition 870250101989, dated 07 / 11 / 2025, page 38 / 122 35 / 87 detection mechanisms are not positioned to capture that part of the environment (e.g., under the tractor). There may be inherent blind spots based on the positioning and orientation of the detection mechanisms 310. For example, blind spot 325 is a part of the surrounding environment that is not in the field of view of any detection mechanism 310. Other blind spots may arise from occlusion by one or more objects in the environment, for example, the ground behind obstacle 320 is hidden from view. V. VERIFICATION OF THE ABSENCE OF OBSTACLES

[0109] In one or more modes, the mowing vehicle can be configured to begin autonomous landscaping actions only if a mowing vehicle manager checks the environment. That is, a mowing vehicle manager must check for obstacles in areas with and without obstacles around the mowing vehicle. Essentially, the manager must walk around the mowing vehicle to check for obstacles in areas undetectable by the detection system. In some configurations, the verification process may include triggering sirens and flashing lights to make it apparent that the mowing vehicle is about to begin autonomous construction. The lights and sirens increase the likelihood that anyone in the environment will leave the area.

[0110] As part of the verification process, the cutting vehicle can communicate with a control system operated by the manager. That is, the cutting vehicle can transmit and receive information from a control system operated by a manager. For example, the cutting vehicle can transmit a request for the manager to check the environment, and the cutting vehicle can receive a verification of the environment in response (once the manager has checked the environment). VI. GENERATING A VIRTUAL SECURITY BUBBLE Petition 870250101989, dated 07 / 11 / 2025, page 39 / 122 36 / 87

[0111] The cutting vehicle includes a virtual safety bubble generation module configured to generate a virtual safety bubble. A virtual safety bubble is an area in the environment that allows the cutting vehicle to operate autonomously without colliding with obstacles. A virtual safety bubble can be an area in the environment (1) directly around the cutting vehicle, (2) on a path ahead of the cutting vehicle, (3) on a path behind the cutting vehicle, (4) along an expected path of the cutting vehicle and / or some area in the environment.

[0112] The cutting vehicle generates the virtual safety bubble based on the cutting vehicle configuration. Here, “configuration” is a term used to describe various aspects of the cutting vehicle, implement, and environment that can be used to generate the virtual safety bubble. A non-exhaustive list of cutting vehicle configuration aspects that can affect the virtual safety bubble follows.

[0113] Machine Path. The machine path can describe a machine's current path or an expected machine path. The machine path can be in any direction relative to the current position of the cutting vehicle. In addition, the virtual safety bubble for the machine path can take into account the characteristics of the cutting vehicle's machine. For example, the virtual safety bubble for a large cutting vehicle along its path is larger than that of a smaller cutting vehicle.

[0114] Vehicle Type. The vehicle type indicates one among a plurality of different vehicles in operation in the landscape environment.

[0115] Landscape Mechanism Configuration. The landscape mechanism configuration indicates a state of the landscape mechanism. For example, the configuration may indicate the position and / or orientation of one or more parts of the landscape mechanism. The configuration may indicate whether the landscape mechanism is activated or not. Petition 870250101989, dated 07 / 11 / 2025, page 40 / 122 37 / 87

[0116] Speed. The speed can be the actual or programmed speed of the cutting vehicle. As implemented by the cutting vehicle, the speed can be a scalar or a vector.

[0117] Acceleration. Acceleration can be an actual or programmed addition of the cutting vehicle. As implemented by the cutting vehicle, acceleration can be scalar or vector.

[0118] Expected obstacle characteristics. The expected obstacle characteristics are obstacle characteristics that a cutting vehicle can expect to encounter in its environment. For example, a cutting vehicle operating near a building can expect to encounter different obstacles than those operating in a landscaped environment. Thus, each environment may have correspondingly different virtual safety bubbles.

[0119] Implement Type. The implement type is the type of implement used by the cutting vehicle (if any). For example, an implement may be a component for performing a landscaping action that can be attached to and / or detached from the vehicle.

[0120] Mounting Mechanism Type. The mounting mechanism type describes how various parts of the cutting vehicle are attached to the frame. For example, a mounting mechanism might be a coupling, and the coupling might be a movable or static coupling. Consequently, the mounting mechanism type can indicate parameters for the virtual safety bubble.

[0121] Type of Landscaping Actions. Landscaping actions are described in detail above. Different landscaping actions may indicate different parameters for the virtual safety bubble. For example, a virtual safety bubble for mowing the lawn may be different from a virtual safety bubble for fertilizing a landscaped area. The mowing vehicle control system may determine the direction for the Petition 870250101989, dated 07 / 11 / 2025, page 41 / 122 38 / 87 which landscaping action would be geared toward assisting in determining the parameters of the virtual safety bubble (e.g., the shape and size of the virtual safety bubble). For example, the control system might set the shape of the virtual safety bubble so that it is predominantly in front of the cutting vehicle based on the landscaping action. In the lawn mowing example, the control system might set the shape of the virtual safety bubble to take into account the positioning of a ramp that guides the clippings.

[0122] Implementation Characteristics for Landscaping Actions. Implementation characteristics describe the details of how a cutting vehicle implements a landscaping action.

[0123] Machine Characteristics for Cutting Vehicle. The machine characteristics describe the physical manifestation of the cutting vehicle. That is, the size, shape, weight, and spatial characteristics of the cutting vehicle. The cutting vehicle can store a digital representation of its machine characteristics, which can be accessed by generating a virtual security bubble.

[0124] Implement Characteristics for the Implement. Implement characteristics describe the physical manifestation of the construction implement. That is, the size, shape, and spatial characteristics of the construction implement. The construction implement can store a digital representation of its implement characteristics, which can be accessed by generating a virtual security bubble.

[0125] Features of Other Accessories. Other accessories may include any component connected to the cutting vehicle or implement. For example, the cutting vehicle may be fitted with additional spotlights that may expand the dimensional profile of the cutting vehicle.

[0126] Environmental Characteristics. The characteristics of the environment Petition 870250101989, dated 07 / 11 / 2025, page 42 / 122 39 / 87 describe the working environment of the cutting vehicle. Some examples of environmental characteristics include the size, shape, and spatial features of the landscape environment in which the cutting vehicle operates. Environmental characteristics may also describe the climate.

[0127] Obstacle Type. Obstacles can be dynamic (i.e., moving) or static (i.e., stationary). The obstacle type can be further classified, for example, between humans or non-humans, between construction equipment, etc. The cutting vehicle can generate a different virtual safety bubble for an identified dynamic and / or static obstacle.

[0128] Manager Input. Manager input is manager information that can be used to generate a virtual security bubble. Manager input can include any of the configuration information mentioned earlier.

[0129] Local Regulations. The control system may keep a record of different local regulations depending on the geographical location of the cutting vehicle. In one or more examples, a first country may have different regulations from a second country; a first state may have different regulations from a second state; a first city may have different regulations from a second city; or some combination thereof. The different regulations may limit landscaping actions, for example, speed limit, permitted operating period, permitted weather conditions for operation, other regulations, etc.

[0130] To update, the cutting vehicle uses a machine setting to determine a virtual safety bubble around the cutting vehicle. The machine setting can be any of the configuration information mentioned previously. The virtual safety bubble can be represented as a relative distance, a Petition 870250101989, dated 07 / 11 / 2025, page 43 / 122 40 / 87 absolute distance, a depth, a time (e.g., based on speed and / or acceleration), legal requirements, or any other metric suitable for quantifying the virtual security bubble.

[0131] The cutting vehicle continuously monitors the environment to ensure there are no obstacles within the virtual safety bubble. That is, the detection mechanisms capture images, the cutting vehicle applies an obstacle identification model to the images, and identifies and locates an obstacle in the environment. If the cutting vehicle identifies an obstacle within the virtual safety bubble, it implements responsive measures to ensure the safe operation of the cutting vehicle.

[0132] Notably, the cutting vehicle can treat obstacles and objects in different ways. For example, a cutting vehicle might identify a large pile of leaves in a virtual safety bubble, identify it as an object, and continue performing landscaping actions because the leaves would not damage the cutting vehicle on contact. Conversely, a cutting vehicle might identify a log in a virtual safety bubble, identify it as an object, classify it as an obstacle, and stop landscaping actions because the log would damage the cutting vehicle in a collision.

[0133] In some instances, the cutting vehicle may treat different types of obstacles differently. For example, a dynamic obstacle (e.g., a human, a moving car, etc.) may warrant different virtual safety bubbles compared to a static obstacle (e.g., a log, a chair, etc.). Naturally, dynamic obstacles are likely to warrant larger virtual safety bubbles due to their ability to move through the environment, while static obstacles are likely to warrant smaller virtual safety bubbles because they remain stationary. In some instances, the cutting vehicle may treat humans differently from all others. Petition 870250101989, dated 07 / 11 / 2025, page 44 / 122 41 / 87 obstacles. For example, the cutting vehicle can generate a virtual safety bubble for humans that is larger than all other objects and obstacles. In one or more modes, the cutting vehicle can generate a plurality of virtual safety bubbles used simultaneously. A first virtual safety bubble can be defined for a first class of objects (e.g., humans) and a second virtual safety bubble can be defined for a second class of objects (e.g., obstacles).

[0134] FIGS. 4A to 4D illustrate different configurations of exemplary safety bubbles.

[0135] FIG. 4A illustrates a first virtual safety bubble around a cutting vehicle, according to one or more embodiments. In one example of a bubble, the cutting vehicle 400 can generate and utilize a single virtual safety bubble 410. As the operation changes, the configuration of the cutting vehicle 400 also changes. Based on the altered configuration, the control system can modify the dimensionality of the virtual safety bubble 410. The cutting vehicle 400 can detect the proximity of the obstacle 490 to the virtual safety bubble 410. If it is within a proximity limit, the cutting vehicle 400 can begin to implement response measures. If the virtual safety bubble 410 is breached, the cutting vehicle 400 can implement additional response measures.

[0136] FIG. 4B illustrates two virtual safety bubbles around the cutting vehicle 400, according to one or more modes. In this mode, the cutting vehicle 400 generates two virtual safety bubbles 410 and 420, with different logics for the objects that interact with each virtual safety bubble. For example, if obstacle 490 violates the first virtual safety bubble 410 (the outermost one), the cutting vehicle 400 can generate alerts. If obstacle 490 violates the second virtual safety bubble of Petition 870250101989, dated 07 / 11 / 2025, page 45 / 122 42 / 87 virtual safety 420, the cutting vehicle 400 may terminate or otherwise modify the operation of the cutting vehicle 400 (e.g., to bypass obstacle 490, to turn off a landscaping mechanism, etc.).

[0137] FIG. 4C illustrates three virtual safety bubbles around the cutting vehicle 400, according to one or more modes. The cutting vehicle 400 can generate the first virtual safety bubble 410, the second virtual safety bubble 420, and the third virtual safety bubble 430. Each safety bubble can be accompanied by different logic to trigger responsive measures.

[0138] FIG. 4D illustrates the dynamic modification of a virtual safety bubble around the cutting vehicle 400 based on the planned trajectory of the cutting vehicle, according to one or more embodiments. In the embodiment, the cutting vehicle 400 is planned to move to the left. Based on the planned trajectory 440, the cutting vehicle 400 can modify the virtual safety bubble 410 to take into account the direction of the cutting vehicle 400.

[0139] In one or more embodiments, the control system may generate a dynamic safety bubble. The dynamic safety bubble is modified based on the configuration of the cutting vehicle. For example, the control system creates a larger safety bubble for a faster cutting vehicle, compared to a smaller safety bubble for a slower cutting vehicle. In other instances, if the visibility of the detection mechanism is limited, the control system may increase the size of the safety bubble to proceed more safely. In one or more instances, if the visibility of the detection mechanism is severely impaired, the control system may increase the size of the safety bubble to a very large size (e.g., up to an infinite size bubble).

[0140] In one or more forms, depending on the cutting vehicle Petition 870250101989, dated 07 / 11 / 2025, page 46 / 122 43 / 87 activates a landscaping mechanism; the control system can dynamically modify the safety bubble. For example, in a mowing context, the mowing vehicle can increase the size of the virtual safety bubble when the mowing platform is activated to cut grass. The mowing vehicle can also take into account an exit trajectory for plant debris. The control system can consider these positional settings when generating the safety bubble, for example, so that the positioning of the ramp can affect a certain dimensionality of the safety bubble in conjunction with the direction of travel of the mowing vehicle, affecting the dimensionality of the safety bubble.

[0141] In some embodiments, the control system can generate a plurality of safety bubbles for use together. Each safety bubble can have a different size and / or shape. The control system can also follow different logic with each safety bubble. For example, if an obstacle violates a specific safety bubble, autonomous operation can be terminated or paused, while for another safety bubble, an audible notification is presented by the vehicle (e.g., via a loudspeaker) to alert those who may be in the environment around the cutting vehicle 100. In another embodiment, a safety bubble can be accompanied by logic to modify the operation of the cutting vehicle 100 based on an identified object violating the safety bubble.

[0142] In some embodiments, the control system can determine the proximity of an object to a safety bubble. If the object is within a proximity limit, the control system can trigger logic to modify the operation of the cutting vehicle. For example, if there is an object less than one meter from the safety bubble, the control system can decelerate the cutting vehicle. The system of Petition 870250101989, dated 07 / 11 / 2025, page 47 / 122 44 / 87 control can determine the amount of deceleration based on the object's behavior. For example, the control system can determine the object's speed and / or trajectory. Based on the speed and / or trajectory, the control system can control the movement of the cutting vehicle, for example, to prevent the object from colliding with the vehicle and / or breaking a safety bubble.

[0143] In one or more embodiments, the control system may use logic to execute different corrective actions in response to a breached safety bubble. In one or more embodiments, the control system may terminate the autonomous operation of the cutting vehicle, for example, by triggering control signals to decelerate any autonomous movement of the cutting vehicle until it stops. In other embodiments, the control system may identify the behavior of the object or obstacle that breaches the safety bubble in order to trigger corrective actions, for example, by triggering control signals to avoid collisions. In some embodiments, the control system may determine whether the object that breaches the safety bubble was previously authorized to breach it.For example, in a landscaped environment with multiple vehicles operating together, the control system can identify other vehicles from sensor data and mark the identified vehicles with permission to violate the safety bubble without interrupting operation. In such modes, the control system can implement different logic for different classes of objects identified as violating the safety bubble. This can be advantageous, for example, when vehicles pass each other on a common lane or route. VII. EXEMPLARY DYNAMIC SAFETY BUBBLE WORKFLOW

[0144] The cutting vehicle can be configured to generate a bubble. Petition 870250101989, dated 07 / 11 / 2025, page 48 / 122 45 / 87 virtual safety bubble around the cutting vehicle, allowing for the safe and autonomous implementation of landscaping actions. FIG. 5 illustrates a process flow to generate a virtual safety bubble, according to an exemplary embodiment. Although FIG. 5 is described from the perspective of the cutting vehicle, any component of the cutting vehicle can perform one or more steps (e.g., the control system or 210). In other embodiments, there may be additional steps or fewer steps. In other embodiments, the listed steps may occur in a different order.

[0145] To provide context, an autonomous cutting vehicle is configured with a detection system. The detection system may comprise six cameras positioned around the cutting vehicle, which provide the cutting vehicle with a 360-degree view of the landscape environment. Within the field of view, there are obstructed views and unobstructed views. Obstructed views are image data within the field of view where an object in the environment obscures parts of the environment behind the object of the detection mechanism (e.g., behind a tire or under the cab). Unobstructed views are image data within the field of view that are not obstructed.

[0146] The mowing vehicle receives a notification to initiate the autonomous implementation of landscaping actions in the environment. In response, the mowing vehicle transmits a request to verify that the mowing vehicle's operating environment is safe. The verification may include transmitting a notification to the manager to check that there are no obstacles in the obstructed views of the detection system. The manager verifies that there are no obstacles and transmits a notification to the mowing vehicle reflecting the verification.

[0147] The cutting vehicle receives a 510 notification that there are no obstacles in the blind spots of the detection system. The manager can Petition 870250101989, dated 07 / 11 / 2025, page 49 / 122 46 / 87 provide such notification, for example, by means of a graphical user interface (GUI) running in a mobile application.

[0148] The cutting vehicle checks 520 for unobstructed views of the environment using an obstacle detection model. That is, the cutting vehicle captures one or more images of the environment using the detection system and applies an obstacle detection model to the images. The obstacle detection model analyzes the images to determine if any of the pixels in the image represent an obstacle.

[0149] The cutting vehicle receives 530 instructions from the operator to begin autonomously performing landscaping actions in the landscaped environment. In an example configuration, the cutting vehicle may not be able to start autonomous execution without a check from the manager that there are no obstacles in the obstructed views and without checking (itself) that there are no obstacles in the unobstructed views.

[0150] The cutting vehicle determines 540 a cutting vehicle configuration to perform the prescribed landscaping actions in the environment. The configuration determination may include access to an implement's capacity, a computational model of the cutting vehicle, types of landscaping actions, and implementation characteristics that define how the cutting vehicle implements the landscaping actions (e.g., speed, trajectory, etc.).

[0151] The cutting vehicle determines a virtual safety bubble based on the determined configuration. The virtual safety bubble represents an area around the cutting vehicle where, if an obstacle is detected in the area, the cutting vehicle will stop operating. The virtual safety bubble can be a distance, a time, a depth, a relative position, or any other measure of a virtual safety bubble.

[0152] The cutting vehicle detects an obstacle in the environment with 560 Petition 870250101989, dated 07 / 11 / 2025, page 50 / 122 47 / 87 is based on applying the obstacle detection model to images captured by the detection system. As the cutting vehicle performs landscaping actions in the landscaped environment, the detection mechanism continuously captures images of the environment. Furthermore, the cutting vehicle continuously applies the obstacle detection model to the captured images to identify obstacles in the environment.

[0153] The cutting vehicle determines 570 that an obstacle is within the virtual safety bubble. The cutting vehicle can determine that the obstacle has exceeded the virtual safety bubble if the depth of the obstacle is within or below the virtual safety bubble. The depth can be determined by means of a detection and range sensor, or by a depth estimation model applied to the images.

[0154] In response to the determination that an obstacle is in the virtual safety bubble, the cutting vehicle terminates operation 560. That is, the cutting vehicle stops implementing landscaping actions in the landscaped environment. In other modes, the cutting vehicle may implement other preventive measures in response to the detection of an obstacle that has crossed the virtual safety bubble. VIII. EXAMPLES OF INTERACTIONS WITH THE MANAGER

[0155] As described above, the cutting vehicle can interact with a manager when performing landscaping actions in the landscaped environment. Some of these interactions may relate to the moment the cutting vehicle detects an object in its virtual safety bubble. Once detected, the cutting vehicle can transmit or receive information from a cutting vehicle manager. The cutting vehicle can also transmit and receive information by establishing a virtual safety bubble around itself. FIGS. 6 to 10 illustrate several examples of Petition 870250101989, dated 07 / 11 / 2025, page 51 / 122 48 / 87 a client device interacting with a cutting vehicle.

[0156] FIG. 6 illustrates a verification process for the cutting vehicle's detection systems. The verification process may include checking that there are no visible obstacles in obstructed views of the cutting vehicle. In the left panel, the graphical user interface (GUI) illustrates the cutting vehicle and implement with six zones where the cameras are positioned and directed. The GUI prompts the manager to “walk around the machine to validate the cameras.” As the manager physically walks around the cutting vehicle, each of the detection mechanisms (e.g., cameras) can capture data that is used by the cutting vehicle to validate the ability of the detection mechanisms to detect the manager. The right panel shows a complete walk with check marks next to each detection mechanism (e.g., camera).

[0157] FIG. 7 illustrates a notification that the cutting vehicle is establishing the virtual safety bubble. That is, once implemented, the virtual safety bubble will be maintained according to the methods described above. Thus, if a person or object enters the virtual safety bubble, the cutting vehicle can perform the corresponding actions, as described above. The left panel shows a slider 710 that allows a manager to involve the cutting vehicle in landscaping actions. Sliding slider 710 to the right is a way of performing step 530 in FIG. 5 of providing and receiving instructions to begin performing landscaping actions autonomously.

[0158] FIG. 8 illustrates a notification transmitted to a client device about a detected obstacle. The notification can occur when the object is detected within the virtual safety bubble. The notification may include characteristics that describe the detected object. In the panel Petition 870250101989, dated 07 / 11 / 2025, page 52 / 122 49 / 87 left, an obstacle notification 810 is displayed as a pop-up notification on a mobile device. Upon receiving a click from the manager, the mobile application can expand to provide a detailed obstacle report 820, shown in the central panel, with additional details about the detected obstacle. The detailed obstacle report 820 may include an option to access a video feed of obstacles 830 captured by a detection mechanism, shown in the right panel. The detailed obstacle report 820 may also include preventive measures that can be taken by the cutting vehicle.

[0159] FIG. 9 illustrates the actions that the cutting vehicle can implement upon detecting an object in the virtual safety bubble. For example, the cutting vehicle can bypass the object in the landscape environment. The graphical user interface (GUI) can illustrate a route around the obstacle and the progress of the cutting vehicle navigating the route, shown in the left panel. After the route is completed, the graphical user interface (GUI) can notify the manager of the successful obstacle avoidance, shown in the center panel. The right panel illustrates another example screenshot showing an alternative route around an obstacle with an actionable option to instruct the cutting vehicle to perform the preventive measure of bypassing the obstacle.

[0160] FIG. 10 illustrates actions that the cutting vehicle can implement when detecting an object in the virtual safety bubble. For example, the cutting vehicle can interrupt operation in the landscape environment. In the left panel, the GUI illustrates that the cutting vehicle ceased operations (paused) in response to the detection of an obstacle. In the middle panel, the GUI indicates that the cutting vehicle is shutting down after being “idle for 30 minutes” after pausing due to obstacle detection. In the right panel, the GUI indicates that the cutting vehicle is “shutting down,” for example, switching to an inactive state. Petition 870250101989, dated 07 / 11 / 2025, page 53 / 122 50 / 87 IX. EXEMPLARY NAVIGATION WORKFLOW

[0161] FIG. 11 illustrates a navigation workflow 1100 of the cutting vehicle, according to one or more embodiments. The cutting vehicle may implement the control system 210 as described in FIG. 2. In other embodiments, the navigation workflow 1100 may include additional steps, fewer steps, steps in a different order, or some combination thereof. Although the following description is made from the perspective of the control system 210, the cutting vehicle as a whole may also execute the navigation workflow (e.g., through distributed systems, as opposed to a single control system).

[0162] The control system 210 begins by detecting objects in an operational environment of the cutting vehicle. The control system 210 uses a spatial mechanism 1105 that generates an occupation grid 1110. The occupation grid 1110 is a virtual representation of the spatial environment of the cutting vehicle. The control system 210 may also use a route mechanism 1120 that generates an active path 1125 for the cutting vehicle to follow. The control system 210 may also receive GPS coordinates 1130, for example, from a GPS receiver. The control system 210 performs passive mapping 1115, detecting objects 1135 in the environment of the cutting vehicle. The control system 210 performs object tracking 1140, for example, by constantly updating the position of an object relative to the cutting vehicle within the occupation grid 1110.

[0163] In one or more embodiments, the control system 210 can use object tracking 1140 to determine if an object may have entered a blind spot. The control system 210 can track an object present in a plurality of images. By determining that the object has disappeared from view, that is, is no longer present in any of the images, the control system 210 can determine that the object has entered a blind spot. Petition 870250101989, dated 07 / 11 / 2025, page 54 / 122 51 / 87 a blind spot. In other embodiments, control system 210, knowing that an object is likely in a blind spot, may prompt the user to check whether the object has been removed or remains in the blind spot. In response to the user providing input indicating that the object has been released, control system 210 may continue operation 1185. In response to the user providing input indicating that the object remains in the blind spot, control system 210 may redirect. Control system 210 may request additional input from the manager via step 1155.

[0164] Control system 210 detects an obstacle in the active path 1145. As noted, control system 210 can use a virtual safety bubble to detect when obstacles have breached it. In response to detecting that the obstacle has breached it, control system 210 interrupts operations 1150 (or implements other preventive measures). Control system 210 notifies the manager 1155 about the obstacle in the path (for example, as shown in FIGS. 6 to 10). Control system 210 receives input 1160 from the manager, for example, to approve 1165 the object, i.e., to ignore the object as not being an obstacle, allowing the operation 1185 to continue. Otherwise, the manager can provide input to redirect 1175. In response, control system 210 can redirect the path 1180 around the obstacle. Once released, control system 210 can continue landscaping actions 1185.

[0165] In one or more embodiments, the control system 210 can routinely update the bounding boxes of the objects. The control system 210 can routinely evaluate whether a bounding box for an object is precisely defined for the object. If it is not precisely defined, the control system 210 can implement the Verification Service 1194 to produce bounding boxes 1196 corrected for Petition 870250101989, dated 07 / 11 / 2025, page 55 / 122 52 / 87 the various objects. Having precise bounding boxes increases the accuracy of detection, i.e., when object detection violates the virtual safety bubble. X. EXEMPLARY NAVIGATION SCENARIOS

[0166] FIG. 12 illustrates the navigation of a cutting vehicle 1210 on a straight path 1230, according to one or more modes. The cutting vehicle 1210 is a mode of the cutting vehicle 100 comprising the control system 210. The cutting vehicle 1210 generates the virtual safety bubble 1220 to assist the navigation of the cutting vehicle 1210. As the cutting vehicle 1210 is traveling on the path 1230 and encounters an obstacle 1240 (i.e., the obstacle 1240 breaks the virtual safety bubble 1220), the cutting vehicle 1210 may interrupt operations and / or implement other preventive measures.

[0167] FIG. 13 illustrates the navigation of a cutting vehicle 1310 off-path on a straight path 1330, according to one or more modes. The cutting vehicle 1310 is a version of the cutting vehicle 100, comprising the control system 210. The cutting vehicle 1310 generates the virtual safety bubble 1320 to assist the navigation of the cutting vehicle 1310. In this scenario, the cutting vehicle 1310 is significantly off-path. If the cutting vehicle 1310 determines that it is off-path, it can generate course correction navigation instructions to direct the cutting vehicle 1310 back to the path 1330. The cutting vehicle 1310 can also interrupt operations and / or provide a notification to a manager indicating that the cutting vehicle 1310 is off-path, requesting subsequent instructions.Even outside the trajectory, if the cutting vehicle 1310 encounters an obstacle 1340 (i.e., if the obstacle 1340 breaches the virtual safety bubble 1320), the cutting vehicle 1310 may interrupt operations and / or adopt other preventive measures. Petition 870250101989, dated 07 / 11 / 2025, page 56 / 122 53 / 87

[0168] FIG. 14 illustrates the navigation of a cutting vehicle 1410 on and off-center along a straight path 1430, according to one or more embodiments. The cutting vehicle 1410 is an embodiment of the cutting vehicle 100 comprising the control system 210. The cutting vehicle 1410 generates the virtual safety bubble 1420 to assist the navigation of the cutting vehicle 1410. In this scenario, the cutting vehicle 1410 is on the path, but off-center. If cutting vehicle 1410 determines that it is off-center, cutting vehicle 1410 may generate course correction navigation instructions to steer cutting vehicle 1410 back to the center of trajectory 1430. Even off-track, if cutting vehicle 1410 encounters an obstacle 1440 (i.e., if obstacle 1340 breaches the virtual safety bubble 1420), cutting vehicle 1410 may interrupt operations and / or implement other preventive measures.

[0169] FIG. 15A illustrates the navigation of a cutting vehicle 1510 off-trajectory, but perceived as being on-trajectory, according to one or more modes. The cutting vehicle 1510 is a representation of the cutting vehicle 100, comprising the control system 210. The cutting vehicle 1510 can receive GPS coordinates such that a perceived position 1515 of the cutting vehicle 1510 is on-trajectory, i.e., on-trajectory 1530. However, in reality, the cutting vehicle 1510 is off-trajectory. The cutting vehicle 1510 uses the virtual safety bubble 1520, but will only trigger preventive measures when the obstacle 1540 (which is off-trajectory) enters the virtual safety bubble 1520. Obstacles that are on the actual-trajectory 1530 may not break the virtual safety bubble 1520, so the cutting vehicle will continue operating.In some modes, the 1510 cutting vehicle may receive corrected GPS coordinates, locating the 1510 cutting vehicle off the trajectory, even though it was previously perceived as being on the trajectory. At this point, the 1510 cutting vehicle can generate and execute the correction navigation. Petition 870250101989, dated 07 / 11 / 2025, page 57 / 122 54 / 87 of the course to guide the cutting vehicle 1510 back to trajectory 1530.

[0170] FIG. 15B illustrates the navigation of a cutting vehicle 1510 when on the trajectory, but perceived as being off the trajectory, according to one or more modes. This scenario is the reverse of the scenario in FIG. 15A. If the cutting vehicle 1510 encounters obstacle 1550 on trajectory 1530, although perceived as being off the trajectory, for example, if the perceived obstacle 1555 is not on trajectory 1530, the cutting vehicle 1510 will execute preventive measures.

[0171] FIG. 16 illustrates the navigation of a cutting vehicle 1610 during a turn on a curved path 1630, according to one or more embodiments. Turn-on refers to the curvature perceived by the control system, which corresponds to the target turning curve, to remain on the curved path 1630 during the turn. Turn deviation refers to the turning curve perceived by the control system, rotationally offset from the target turning curve to remain on the curved path 1630. The cutting vehicle 1610 is an embodiment of the cutting vehicle 100 comprising the control system 210. When on a curved path 1630, the cutting vehicle 1610 can adjust the virtual safety bubble 1620 to account for the turning radius of the cutting vehicle 1610. For example, the virtual safety bubble 1620 can be extended in the turning direction of the cutting vehicle 1610.When cutting vehicle 1610 detects one or more obstacles 1640 and 1650 within the virtual safety bubble 1620, cutting vehicle 1610 can implement preventive measures.

[0172] FIG. 17 illustrates the navigation of a cutting vehicle when deviating from a curve in a curved trajectory 1730, according to one or more embodiments. The cutting vehicle 1710 is an embodiment of the cutting vehicle 100, comprising the control system 210. The cutting vehicle 1710 may have a perceived orientation that is distorted in relation to Petition 870250101989, dated 07 / 11 / 2025, page 58 / 122 55 / 87 real orientation. In this scenario, the cutting vehicle 1710 is moving along a perceived curved path 1735 that is offset from the curved path 1730. The cutting vehicle 1710 can implement course correction navigation to align the orientation of the cutting vehicle 1710, that is, to align the perceived path 1735 with the real path 1730. In one or more modes, the cutting vehicle 1710 can use the detection mechanisms to locate obstacles 1740 and 1750 on the path 1730 as markers on the path 1730. XI. SAFETY BUBBLE WORKFLOW FOR CUTTING

[0173] FIG. 18 illustrates a process flowchart describing collision avoidance with 180° object recognition by an autonomous cutting vehicle, according to one or more embodiments. A control system for an autonomous cutting vehicle (e.g., control system 210 of FIG. 2) can perform collision avoidance with 180° object recognition in conjunction with other components of the autonomous cutting vehicle. In other embodiments, one or more steps may be performed by other systems or devices in conjunction with the autonomous cutting vehicle.

[0174] The autonomous cutting vehicle establishes one or more virtual safety bubbles around the autonomous cutting vehicle to prevent collisions during autonomous operation. In some embodiments, the establishment of the virtual safety bubble(s) includes determining a configuration of the autonomous cutting vehicle. The configuration of the autonomous cutting vehicle can be defined according to a planned field operation. The planned field operation may include a series of actions to be performed by the autonomous cutting vehicle, for example, one action may be to traverse an environment to a landscaping site, another action may be to mow the grass of the landscaping site. Petition 870250101989, dated 07 / 11 / 2025, page 59 / 122 56 / 87 etc. For each action, the control system can configure the autonomous cutting vehicle distinctly to perform that action. For example, when traversing an environment, the control system can identify an ideal path or route to traverse the environment. In another example, when mowing the lawn of a landscaping site, the control system can identify an ideal path to efficiently mow the lawn, i.e., minimize redundant passes over cut areas. The control system can establish different sets of one or more safety bubbles for each configuration. For example, when traversing landscapes, the control system can utilize at least one virtual safety bubble to avoid collisions with objects in the environment.In another example, during the execution of the cutting action, the control system may utilize multiple safety bubbles, with one safety bubble to alert any individual within the safety bubble and another safety bubble to terminate the operation of the cutting platform. Other factors affecting the generation of virtual safety bubbles are described throughout this disclosure.

[0175] In one or more embodiments, the establishment of virtual safety bubbles may still utilize a human operator to initiate autonomy. In one or more embodiments, the autonomous cutting vehicle may include a platform for a human operator to stand on the vehicle. The vehicle may include controls to operate in a manual configuration, i.e., with the human operator controlling the operation from the platform, or an autonomous configuration, i.e., without the need for the human operator to control the vehicle from the platform. In some embodiments, to initiate autonomy, the human operator may initiate autonomy from the vehicle's controls. Once initiated, the vehicle may use a timer for the human operator to move beyond one or more established safety bubbles before initiating operations. In other embodiments, to initiate autonomy, the operator Petition 870250101989, dated 07 / 11 / 2025, pp. 60 / 122 57 / 87 A human can initiate autonomy through a client device communicating with the autonomous cutting vehicle. In response to receiving a signal from the client device to initiate autonomy, the autonomous cutting vehicle can establish one or more safety bubbles and assess whether any object is within the established safety bubbles before initiating operation.

[0176] In one or more modes, during autonomous start-up, the autonomous cutting vehicle may monitor the operator's position during an activation timer. In such modes, the operator may activate a switch or control on the autonomous cutting vehicle to activate autonomous mode. Once activated, the operator moves away to release one or more established virtual safety bubbles. If the operator fails to exit one or more virtual safety bubbles, i.e., if the operator's position remains within one or more virtual safety bubbles, depending on which bubble is breached, the autonomous cutting vehicle may implement corrective actions to effect the operator's movement to release the breach. If the operator does not exit the virtual safety bubble before the activation timer expires, the autonomous cutting vehicle cannot initiate autonomous operation.If the operator releases the virtual safety bubble before it expires, the autonomous mowing vehicle can begin autonomous operation. In some modes, the autonomous mowing vehicle can begin autonomous operation by determining that the operator is outside an inner bubble configured to trigger a landscaping mechanism, even though the operator is inside an outer bubble configured to alert bystanders.

[0177] The autonomous cutting vehicle captures 1820 images from a camera system mounted on the autonomous cutting vehicle. The camera system captures images with a view of the environment surrounding the autonomous cutting vehicle. In one or more modes, the camera system Petition 870250101989, dated 07 / 11 / 2025, page 61 / 122 58 / 87 may include stereoscopic camera pairs, for example, a pair of forward-facing cameras, a pair of rear-facing cameras, a pair of left-facing cameras, and a pair of right-facing cameras (as shown in FIG. 3). Adjacent camera pairs may have overlapping fields of view, providing a complete 360-degree panoramic view of the environment surrounding the autonomous cutting vehicle. With the captured image data, the control system can perform one or more image analyses to identify features from the image data. For example, the control system may apply a depth estimation model to the image data to predict the depth of one or more images from the image data. In another example, the control system may apply a semantic segmentation model to classify pixels as one of a plurality of semantic labels (e.g., ground, sky, foliage, obstacle, etc.).In another example, the control system can apply a visual odometry algorithm to the captured image data to determine the object's speed and / or acceleration.

[0178] In some embodiments, the autonomous cutting vehicle may capture additional data from one or more sensors implemented on the vehicle. For example, the autonomous cutting vehicle may include one or more range and detection sensors (e.g., light detection and ranging (LiDAR), radio detection and ranging (RADAR), etc.) to measure depth information of objects in the environment. In embodiments with depth data, the control system may merge depth data from distinct sources to form more accurate measurements. For example, the control system may merge depth information determined from captured image data with depth information from LiDAR data. The fusion process may take into account a weighted average of the depth measurements from the two sources. Petition 870250101989, dated 07 / 11 / 2025, page 62 / 122 59 / 87

[0179] The autonomous cutting vehicle generates a spatial representation of the environment from the captured images. The autonomous cutting vehicle can generate a three-dimensional spatial representation of objects in the environment. The autonomous cutting vehicle can project pixels from the captured images into 3D space based on depth information (e.g., determined by the captured images or LiDAR data). The autonomous cutting vehicle can group points to identify multiple objects in the environment. For example, objects can be identified using an object detection algorithm applied to the captured image data. Pixels are grouped and labeled as components of an object. Points projected from the grouped pixels can also be grouped in the spatial representation to form the object in the spatial representation.In one or more modes, the autonomous cutting vehicle can combine one or more polygons to form the object's boundary shape in the spatial representation.

[0180] The autonomous cutting vehicle performs 1840 object detection and characterization from spatial representation. The autonomous cutting vehicle can use one or more image-based models to detect and characterize objects. For example, a machine learning model can be used to classify each object identified in the spatial representation. Object classes can be broad categories (e.g., person, vehicle, animal, inanimate object, etc.). One or more classes can have subcategories. For example, people can be subdivided into roles such as observer or operator. In another example, vehicles can be subdivided into different types, e.g., self-propelled, motorized, large, small, etc. The model can also identify specific instances, e.g., Person A is identified in the environment, Vehicle B is identified, etc. One or more models can also determine object characteristics, e.g., Petition 870250101989, dated 07 / 11 / 2025, page 63 / 122 60 / 87 speed, trajectory, acceleration / deceleration, dimension(s), color, perception (assuming an animated object, such as a person or an animal), behavior, proximity, etc. The models can use image data, spatial representation, other sensor data, or some combination of these as input.

[0181] The autonomous cutting vehicle generates a virtual buffer for each object based on the object's characteristics. The autonomous cutting vehicle can determine the dimensionality of the virtual buffer based on the object's characteristics. For example, the virtual buffer can be large, medium, or small. The large buffer, for example, can have four times the dimensionality of the object. The medium buffer, for example, can have three times the dimensionality of the object. The small buffer, for example, can have twice the dimensionality of the object. Any number of buffer sizes or dimensionalities can be implemented. The sizing of the virtual buffers can also be determined based on the vehicle size, for example, a larger vehicle has a greater stopping distance, so the virtual buffer is sized for a smaller vehicle.

[0182] FIG. 19 illustrates an example of a decision tree for determining the size of the virtual buffer based on the object's characteristics. In an initial assessment, the autonomous cutting vehicle determines 1900 whether the object is a moving obstacle. If so, the autonomous cutting vehicle determines 1910 whether the object has the ability to move on its own. If so, the autonomous cutting vehicle may determine 1920 that the object needs a large buffer (e.g., people or animals). If the autonomous cutting vehicle determines that the object cannot move on its own, the autonomous cutting vehicle determines 1930 a medium buffer for the object (e.g., vehicle). If the autonomous cutting vehicle determines that the object cannot move, the autonomous cutting vehicle determines 1940 whether the object has the ability to cause damage to Petition 870250101989, dated 07 / 11 / 2025, page 64 / 122 61 / 87 objects around. If so, the autonomous cutting vehicle determines a medium buffer of 1930 for the object. Otherwise, the autonomous cutting vehicle determines a small buffer of 1950 for the object. In some embodiments, the dimensionality of each buffer size is objective. For example, the large buffer has a radius of 2 meters around the object. In other embodiments, the dimensionality may depend on the dimensionality of the object, for example, the large buffer is four times the size of the object. So an object with an area of ​​1 meter by 1 meter may have a large buffer with an area of ​​4 meters by 4 meters, centered around the object.

[0183] Returning to FIG. 18, in one or more embodiments, the autonomous cutting vehicle can generate a user interface for displaying the status of the autonomous cutting vehicle and for annotating objects detected in the environment. The user interface can render the spatial representation of the environment around the autonomous cutting vehicle, including the objects identified in the environment. The user interface can provide dynamic real-time feed-in to an operator's client device, providing the operator with a real-time view of the autonomous cutting vehicle's operation. The user interface can further annotate objects in the environment based on analyses performed by the autonomous cutting vehicle. For example, objects can be annotated based on their classifications (e.g., person, animal, vehicle, etc.).The user interface can also display the virtual buffer generated for each object (e.g., small buffer, medium buffer, large buffer, etc.). The user interface can annotate the object with other characteristics, such as speed, perception, behavior, etc.

[0184] The autonomous cutting vehicle operates autonomously, including evading the violation of the virtual security bubble by virtual buffers of any objects in the operating environment. The autonomous vehicle Petition 870250101989, dated 07 / 11 / 2025, page 65 / 122 The 62 / 87 cutting vehicle can perform landscaping operations, traverse the environment, etc., based on a landscaping plan designed for the autonomous cutting vehicle. The autonomous cutting vehicle continues the autonomous operation carefully to avoid violating the virtual safety bubble of the autonomous cutting vehicle. Evading the violation of the virtual safety bubble(s) involves avoiding the collision of the virtual protection zone of any object with the virtual safety bubble(s). The vehicle can project a vehicle trajectory and / or an object trajectory to detect a potential collision if the vehicle and / or the object continue on their respective trajectories. The vehicle can also incorporate some tolerance in the evasive procedure, for example, a change of direction action to prevent the virtual protection zone of an object from approaching a limit distance (e.g., 1 meter) from the virtual safety bubble.In some embodiments, the autonomous cutting vehicle can perform real-time corrective actions to avoid the violation. For example, the autonomous cutting vehicle can determine a safe alternate route to navigate to and avoid an object. In another example, the autonomous cutting vehicle can use notifications or other warnings to effect changes in the object's behavior. In one or more examples, the autonomous cutting vehicle can modify the operation or terminate the autonomy. Other actions can be performed in conjunction with the virtual safety bubble and virtual protection zones for the objects.

[0185] In one or more exemplary implementations, the autonomous cutting vehicle can perform object evasion based on the autonomous cutting vehicle's operating mode, the classification of a nearby object, the characterization of the nearby object, or some combination thereof.The operating mode of the autonomous cutting vehicle may include, for example, a stationary mode (i.e., the autonomous cutting vehicle is stationary with no actively engaged mechanisms), a mode of... Petition 870250101989, dated 07 / 11 / 2025, page 66 / 122 63 / 87 transit (i.e., the autonomous cutting vehicle is in motion, but without actively activated mechanisms), a landscaping mode (i.e., the autonomous cutting vehicle is in motion and at least one mechanism is actively activated in performing a landscaping action), and autonomous startup or shutdown. Object classification can include a person, a vehicle, an operator, an animal, an inanimate object, etc. People can be classified separately into specific roles, e.g., operator, observer, etc. For example, object characterization can include an object's behavior.In one or more examples, the behavior of an animated object (e.g., a person, an animal, or a person-controlled vehicle) may include a curious state (i.e., the object is aware of the presence of the autonomous cutting vehicle and observes it), an oblivious state (i.e., the object is unaware of the presence of the autonomous cutting vehicle), an evasive state (i.e., the object is dodging the autonomous cutting vehicle), an aggressive state (i.e., the object is approaching the autonomous cutting vehicle), and a trained state (i.e., an operator is aware of the presence of the autonomous cutting vehicle and interacts with the autonomous cutting vehicle).

[0186] The autonomous cutting vehicle can classify a person’s behavior based on their movement relative to the autonomous cutting vehicle. For example, a person who breaks through an external virtual safety bubble and then moves away from the autonomous cutting vehicle can be classified as being in an evasive state. With people, the autonomous cutting vehicle can operate with increased sensitivity. For example, the autonomous cutting vehicle can increase the size of a virtual safety bubble that triggers the termination of autonomous operation with a person compared to a non-human object.

[0187] In one or more forms, the autonomous cutting vehicle Petition 870250101989, dated 07 / 11 / 2025, page 67 / 122 64 / 87 can generate an external virtual safety bubble to trigger warnings that cause animated objects to move away from the vehicle. The external virtual safety bubble can be the largest virtual safety bubble. In case of a violation, the vehicle can trigger warnings (including an audible signal through a loudspeaker and / or a visual signal through a lighting system). The vehicle can also start a timer. If the object does not leave the external virtual safety bubble, the vehicle can terminate operation or interrupt the activation of a landscaping mechanism, which may have been triggered at the time of the violation.

[0188] In one or more embodiments, the autonomous cutting vehicle may generate an internal virtual safety bubble, for example, when the landscaping mechanism is activated. The internal virtual safety bubble may be used in conjunction with one or more other virtual safety bubbles. Upon violation of the virtual buffer of any object, the autonomous cutting vehicle may interrupt the operation of the landscaping mechanism.

[0189] In one or more embodiments, to avoid the violation of virtual safety bubbles by the virtual buffer of any object, the autonomous cutting vehicle can modify the autonomous cutting vehicle configuration. For example, the vehicle can detect a vehicle trajectory and, optionally, an object trajectory (for a moving object). The vehicle can determine a potential collision based on the vehicle trajectory and, optionally, the object trajectory. The vehicle can determine how to modify the vehicle configuration that would prevent the violation, for example, by slowing down, turning, etc. The vehicle modifies the configuration, thereby modifying the virtual safety bubbles. For example, slowing down the vehicle can result in the shrinking of the virtual safety bubble. As another example, ending the landscaping mechanism's operation can result in the removal of a virtual safety bubble.

[0190] In one or more forms, the autonomous cutting vehicle Petition 870250101989, dated 07 / 11 / 2025, p. 68 / 122 65 / 87 can update the 1880 classification and / or virtual buffers for objects based on monitored behavior. As the autonomous cutting vehicle operates autonomously, it continues to collect data that describes and reflects the environment around the autonomous cutting vehicle. In some modes, as the autonomous cutting vehicle takes corrective actions to avoid violation, it continues to monitor the behavior of objects in the environment. Based on this monitoring, the autonomous cutting vehicle can update the classification of objects. For example, the autonomous cutting vehicle can increase the aggressiveness level in classifying an object's behavior based on the object's persistent movement toward the autonomous cutting vehicle. The autonomous cutting vehicle can also update the virtual buffer for an object based on the updated classification and / or characterization.Using another example, the virtual buffer can be reduced by determining that a moving object has stopped moving.

[0191] The autonomous mowing vehicle continues to operate autonomously until certain stopping conditions are met. In one or more modes, the autonomous mowing vehicle may complete the landscaping plan by executing the landscaping actions included in the plan. Once completed, the autonomous mowing vehicle may terminate autonomous operation. In one or more modes, the autonomous mowing vehicle may terminate autonomous operation in the event of a breach of the virtual safety bubble. In such examples, the autonomous mowing vehicle may transmit a notification to the operator about the breach and the termination of operation.

[0192] FIG. 20 illustrates a workflow for implementing virtual safety bubbles for a human detected near the cutting vehicle, according to one or more modalities.

[0193] The workflow can be executed by a system of Petition 870250101989, dated 07 / 11 / 2025, page 69 / 122 66 / 87 control of a cutting vehicle using three or more virtual safety bubbles. A warning bubble can be the largest bubble, to alert pedestrians or other living beings in transit. A collision bubble can be smaller than the warning bubble and used to identify potential collisions with the cutting vehicle. A hazard bubble can be smaller than the warning bubble (and optionally smaller than the collision bubble) and used to identify potential hazards to humans.

[0194] A human candidate detection 2000 is identified by sensor data by the detection mechanisms (e.g., camera image data). The control system determines 2002 whether the human candidate has crossed the hazard bubble. If so, the control system triggers an emergency stop 2004. The control system continues to generate and transmit a notification to the operator 2006 about the detection of the human candidate. The control system may also start a timer 2014 used to reset the autonomy, assuming the safety conditions are met. The control system implements a human verification model 2008 (described in more detail in FIG. 22) to verify whether the detected object is indeed a human. If the human verification model 2008 verifies the presence of a human within the hazard bubble, the control system terminates the autonomy 2010.If the 2008 human verification model rejects the human classification, the control system can resume the autonomy of the cutting vehicle. The results of the 2008 human verification model can be recorded, for example, for human review or to refine the control system models. If the model determines that the object is not human, the control system can enact a workflow for non-human objects (e.g., described in FIG. 21).

[0195] During and until the end of the timer, the control system evaluates whether the human candidate has crossed the risk bubble. In case Petition 870250101989, dated 07 / 11 / 2025, pp. 70 / 122 67 / 87 affirmative, the control system can resume autonomy in 2012. Otherwise, the control system can end autonomy in 2010.

[0196] If the control system determines that the human candidate is not in the risk bubble, the control system determines 2018 whether the human candidate is in the collision bubble. If so, the control system reduces the speed 2020. The control system further tracks 2022 the human candidate. If the speed has not yet been reduced to 0 mph, the control system may further reduce the speed 2024. The control system may also initiate another timer 2026. The control system may use the human verification model 2008 to verify whether the human candidate is indeed human. If verified, the control system may assess 2028 whether the human has crossed the collision bubble. Otherwise, the control system may terminate the autonomy 2010. If, at any time, the human candidate object has crossed the collision bubble, the control system may continue the operation 2034.

[0197] If the control system determines that the candidate human object is not in the collision bubble, the control system determines 2030 whether the candidate human object is in the warning bubble. If so, the control system may generate warning alerts 2032 to notify the candidate human object to move away from the operating environment. Warning alerts may include any type of output perceptible to a human, for example, audio, visual, client device notification, etc. If the candidate human object is near the warning bubble but has not yet crossed it, the control system may continue operation 2034. Warning alerts may be audible signals (e.g., by a loudspeaker system) or visual signals (e.g., by a lighting system or an electronic display).In other modes, other warning alerts may be used, any combination of warning alerts may be used, or a layered alert scheme may be used. Petition 870250101989, dated 07 / 11 / 2025, pp. 71 / 122 68 / 87

[0198] FIG. 21 illustrates a workflow for implementing a virtual safety bubble to resolve a non-human obstacle detected near the cutting vehicle, according to one or more embodiments. The workflow can be executed by a cutting vehicle control system using at least one virtual safety bubble. The control system can execute this workflow upon detection of a static non-human object 2100. In other embodiments, any combination of the steps described can be incorporated into the workflow. For example, there may be additional steps, fewer in number, or different from those described.

[0199] The control system determines 2102 whether the path is obstructed by the object. In that case, the control system triggers a controlled stop 2104. The control system applies the verification model 2106 to the object (i.e., to the object's image data) to classify it.

[0200] If the verification model 2106 determines that the object is an obstacle, the control system generates and transmits a notification to the operator 2108. The notification may present data about the identified obstacle. The notification may request instructions 2110 on whether the obstacle should be bypassed. If the operator provides instructions to bypass the obstacle, the control system updates 2112 the route to find a safe route around the static obstacle. If the operator does not provide such instructions to bypass the obstacle, the control system detects 2116 whether the obstacle has been removed. Otherwise, the control system terminates the autonomy 2118. If the obstacle is removed, the control system continues the operation 2114.

[0201] If the verification model 2106 identifies the object to be landscaped, the control system determines 2120 whether the cutting vehicle can be bypassed by the landscaping. If so, the control system updates the route 2122. The control system creates a record of the event 2124. Petition 870250101989, dated 07 / 11 / 2025, page 72 / 122 69 / 87 The control system continues operations 2126. After the completion of task 2128, the control system sends the long event to the operator 2130. The operator can then provide updated plans, which are received 2132 by the control system. If the control system determines that the landscaping cannot be bypassed, the control system generates and transmits a notification to the operator 2134 about the obstructed path. The control system ends autonomy 2136.

[0202] The verification model 2106 can further record the results 2138 for human review. Based on the human review, the results can be used to adjust any of the models used by the control system.

[0203] FIG. 22 illustrates a workflow for human verification with multiple human detection concordance, according to one or more modalities. In one or more modalities, a model implements the illustrated workflow (i.e., a human verification model). The workflow is executed by a control system (e.g., control system 130 or control system 210). In other modalities, one or more steps are executed by other systems (e.g., external system(s) 220).

[0204] The control system detects an object on path 2200. The control system performs classification in vehicle 2202 to determine a classification of the object (e.g., human, non-human obstacle, etc.). For example, the classification in vehicle 2202 may use a first type of model or algorithm to classify the object based on data collected from the sensor. The control system then transmits the data to an external system 2206 for a second classification. The control system receives the secondary classification from the external system 2208. The external system 2208 may employ one or more other models and / or algorithms (e.g., which may be trained more robustly than Petition 870250101989, dated 07 / 11 / 2025, page 73 / 122 70 / 87 (which is the model in the vehicle) to classify the object.

[0205] The control system evaluates whether the classification in the vehicle matches the secondary classification of the external system 2210. If the two classifications match, the control system evaluates whether the corresponding classification matches the human label 2212. If a true positive human is detected by both classifications 2214, the control system may terminate the autonomy 2216. If a true positive non-human is detected by both classifications 2218, the control system may notify the operator 2220 to obtain subsequent instructions on how to handle the obstacle. The notification may be sent to a client device in use by the operator, which is also in communication with the vehicle control system.

[0206] The control system receives user input 2222 on how to deal with the obstacle. The user input may indicate that the obstacle will be removed. The control system waits for the obstacle to be removed 2224. The control system may capture additional sensor data to assess whether the object has been removed 2226. If so, the control system may resume autonomy 2228. Alternatively, the user input may indicate that the vehicle should go around the obstacle. Based on this input, the control system navigates around the obstacle 2230. Navigation around the obstacle may include determining a clear route around the obstacle, which resumes the vehicle's path. The control system executes updated navigation to guide the vehicle around the obstacle. After the return route, the control system resumes autonomy 2228.The user input may alternatively indicate that the vehicle should ignore obstacle 2232, for example, if the obstacle is too small, if the obstacle was detected as a false positive, etc. The control system resumes autonomy 2228. The control system may also use the operator override classification as one. Petition 870250101989, dated 07 / 11 / 2025, pp. 74 / 122 71 / 87 Example of feedback training. The control system (or an external system) can retrain the object classification model based on the operator's corrected classification. The control system can also update an environment map to reflect the detected obstacle and its classification.

[0207] If the two classifications (the classification in the vehicle and the classification of the external system) are incompatible, the control system transmits the data to the operator's client device for review by the operator 2234. The data may include images of the object detected in the path. The operator provides user input 2236 on the classification. The control system may receive a corrected classification from the operator, replacing the other two classifications. The control system then begins to evaluate whether the classification is human 2212. Otherwise, the control system may receive input to ignore the classification, so that the control system ignores the obstacle 2232 and resumes autonomy 2228. XII. EXEMPLARY COMPUTING SYSTEM

[0208] FIG. 23 is a block diagram illustrating the components of an exemplary machine capable of reading instructions from a machine-readable medium and executing them in a processor (or controller). Specifically, FIG. 23 shows a diagrammatic representation of a machine in the exemplary form of a computer system 2300, within which program code (e.g., software or software modules) can be executed to cause the machine to perform one or more of the methodologies discussed herein. The program code may consist of instructions 2324 executable by one or more processors 2302. In alternative embodiments, the machine operates as a stand-alone device or may be connected (e.g., networked) to Petition 870250101989, dated 07 / 11 / 2025, pp. 75 / 122 72 / 87 other machines. In a network deployment, the machine can operate in the capacity of a server machine or a client machine in a server-client network environment, or as a peer machine in a point-to-point (or distributed) network environment.

[0209] The machine can be a server computer, a client computer, a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a mobile phone, a smartphone, a web device, a router, a network switch or bridge, or any machine capable of executing 2324 instructions (sequential or not) that specify actions to be taken by that machine. Furthermore, although only a single machine is illustrated, the term “machine” should also be understood as including any set of machines that, individually or jointly, execute the 2324 instructions to perform one or more of the methodologies discussed herein.

[0210] The exemplary computer system 2300 includes a processor 2302 (for example, a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), one or more application-specific integrated circuits (ASICs), one or more radio frequency integrated circuits (RFICs), or any combination thereof), a main memory 2304, and a static memory 2306, which are configured to communicate with each other via a bus 2308. The computer system 2300 may further include a visual display interface 2310. The visual interface may include a software driver that enables the display of user interfaces on a screen (or display). The visual interface may display user interfaces directly (for example, on the screen) or indirectly on a surface, window, or similar (for example, by means of a visual projection unit).To simplify the discussion, the visual interface can be described as a screen. The 2310 visual interface may include or interact with a touch screen. Petition 870250101989, dated 07 / 11 / 2025, page 76 / 122 73 / 87 at the touch. The 2300 computer system may also include an alphanumeric input device 2312 (for example, a keyboard or touch-screen keyboard), a cursor control device 2314 (for example, a mouse, a trackball, a joystick, a motion sensor, or other pointing instrument), a storage unit 2316, a signal generation device 2318 (for example, a loudspeaker), and a network interface device 2320, which are also configured to communicate via bus 2308.

[0211] The storage unit 2316 includes a machine-readable medium 2322 in which instructions 2324 (e.g., software) incorporating one or more of the methodologies or functions described herein are stored. The instructions 2324 (e.g., software) may also reside wholly or at least partially in main memory 2304 or in the processor 2302 (e.g., in a processor's cache memory) during their execution by the computer system 2300, the main memory 2304 and the processor 2302 also constituting machine-readable media. The instructions 2324 (e.g., software) may be transmitted to or received over a network 190 by means of the network interface device 2320.

[0212] Although the machine-readable medium 2322 is shown in an embodiment example as a single medium, the term “machine-readable medium” should be understood as including a single medium or multiple media (e.g., a centralized or distributed database, or associated caches and servers) capable of storing instructions (e.g., 2324 instructions). The term “machine-readable medium” should also be understood as any medium capable of storing instructions (e.g., 2324 instructions) for execution by the machine and causing the machine to execute one or more of the methodologies disclosed herein. The term “machine-readable medium” includes, among others, data repositories in Petition 870250101989, dated 07 / 11 / 2025, page 77 / 122 74 / 87 forms of solid-state memory, optical media, and magnetic media. XIII. EXEMPLARY MODALITIES

[0213] Clause 1.An autonomous cutting vehicle comprising: a camera system comprising a plurality of cameras positioned around the autonomous cutting vehicle; a cutting platform comprising one or more motorized blades for cutting plants in an environment; and a control system configured to: capture image data from a camera system of an autonomous cutting vehicle; detect an object in an environment around the autonomous cutting vehicle based on the image data; generate a virtual buffer for the object, the virtual buffer positioned around the object; generate a plurality of virtual safety bubbles around the autonomous cutting vehicle based on a configuration of the autonomous cutting vehicle; and execute, through at least the cutting platform, the autonomous operation of the cutting vehicle to perform one or more landscaping actions in the environment, avoiding the violation of the plurality of virtual safety bubbles by the virtual buffer of the object.

[0214] Clause 2. The autonomous cutting vehicle of clause 1, the control system further configured to: generate a spatial representation of the autonomous cutting vehicle based on image data, wherein the spatial representation spatially describes the object in relation to the autonomous cutting vehicle.

[0215] Clause 3. The autonomous vehicle cut from clause 2, the control system further configured to: apply one or more machine learning models to classify an object type from among a plurality of object types for the object based on image data, wherein the virtual buffer for the object is based on the object type classified by one or more machine learning models. Petition 870250101989, dated 07 / 11 / 2025, pp. 78 / 122 75 / 87

[0216] Clause 4. The autonomous cutting vehicle of clause 3, wherein the plurality of object types includes persons and inanimate objects, wherein, in response to the classification of the object as a person, it generates the virtual buffer for the object that has a first size, and wherein, in response to the classification of the object as an inanimate object, it generates the virtual buffer for the object that has a second size smaller than the first size.

[0217] Clause 5. The autonomous cutting vehicle of any of clauses 1 to 4, wherein the control system configured to generate the plurality of virtual safety bubbles based on the configuration of the autonomous cutting vehicle is understood to be configured to: generate an external virtual safety bubble; and, in response to the violation of the external virtual safety bubble, trigger one or more warnings to distance the object from the autonomous cutting vehicle.

[0218] Clause 6. The autonomous vehicle cut off from clause 5, wherein the control system configured to trigger one or more warnings comprises being configured for: the external virtual safety bubble being configured to trigger, as one or more warnings, an audible signal by a loudspeaker or a visual signal by a system of lights.

[0219] Clause 7. The autonomous cutting vehicle of any of clauses 5 to 6, wherein the control system configured to generate the plurality of virtual safety bubbles based on the configuration of the autonomous cutting vehicle is further configured to: in response to the violation of the outer virtual safety bubble, start a timer; and in response to the expiration of the timer, terminate the autonomous operation of the autonomous cutting vehicle.

[0220] Clause 8. The autonomous cutting vehicle of any of clauses 1 to 7, in which the control system is configured to generate the plurality of virtual safety bubbles based on the configuration of Petition 870250101989, dated 07 / 11 / 2025, page 79 / 122 76 / 87 autonomous cutting vehicle comprises being configured to: generate an internal virtual safety bubble during the activation of the autonomous cutting vehicle's cutting platform; and in response to a violation of the internal virtual safety bubble, terminate the activation of the cutting platform.

[0221] Clause 9.The autonomous cutting vehicle of any of clauses 1 to 8, wherein the control system being configured to perform the autonomous operation of the cutting vehicle comprises being configured to: detect a potential violation by the virtual buffer of virtual safety bubble plurality based on a trajectory of the autonomous cutting vehicle; in response to the detection of the potential violation, modify the configuration of the autonomous cutting vehicle to avoid the violation of the virtual safety bubble plurality; and modify the virtual safety bubble plurality around the autonomous cutting vehicle based on the modified configuration of the autonomous cutting vehicle.

[0222] Clause 10. A non-transient, computer-readable storage medium that stores instructions which, when executed by a computer processor, cause the computer processor to perform operations comprising: capturing image data from a camera system of an autonomous cutting vehicle; detecting at least one object in an environment around the autonomous cutting vehicle based on the image data; generating a virtual buffer for the object, the virtual buffer positioned around the object; generating a plurality of virtual safety bubbles around the autonomous cutting vehicle based on a configuration of the autonomous cutting vehicle; and executing the autonomous operation of the cutting vehicle to perform one or more landscaping actions in the environment, avoiding violation of the plurality of virtual safety bubbles by the object's virtual buffer.

[0223] Clause 11. The non-transient, computer-readable storage medium of clause 10, in which the image data is captured. Petition 870250101989, dated 07 / 11 / 2025, pages 80 / 122 77 / 87 involves capturing images from stereoscopic camera pairs positioned around the autonomous cutting vehicle to capture a 360-degree panoramic view of the environment surrounding the autonomous cutting vehicle.

[0224] Clause 12. The non-transient, computer-readable storage medium of any of clauses 10-11, the operations further comprising: generating a spatial representation of the autonomous cutting vehicle based on the image data, wherein the spatial representation spatially describes the object in relation to the autonomous cutting vehicle.

[0225] Clause 13. The non-transient, machine-readable storage medium of clause 12, the operations further comprising: applying one or more machine learning models to classify an object type from among a plurality of object types to the object based on image data, wherein the virtual buffer for the object is based on the object type classified by one or more machine learning models.

[0226] Clause 14. The non-transient, computer-readable storage medium of clause 13, wherein the plurality of object types includes persons and inanimate objects, wherein, in response to the classification of the object as a person, the virtual buffer for the object is generated with a first size, and wherein, in response to the classification of the object as an inanimate object, the virtual buffer for the object is generated with a second size smaller than the first size.

[0227] Clause 15. The non-transient, computer-readable storage medium of any of clauses 10 to 14, wherein the generation of the plurality of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an outer virtual safety bubble, wherein the violation of the outer virtual safety bubble triggers one or more distance warnings from the autonomous cutting vehicle. Petition 870250101989, dated 07 / 11 / 2025, page 81 / 122 78 / 87

[0228] Clause 16. The non-transient, computer-readable storage medium of clause 15, in which the external virtual security bubble is configured to trigger, as one or more warnings, an audible signal by a loudspeaker or a visual signal by a lighting system.

[0229] Clause 17. The non-transient, computer-readable storage medium of any of clauses 15-16, wherein the external virtual safety bubble is configured to, in response to the expiration of a timer, terminate the autonomous operation of the autonomous cutting vehicle.

[0230] Clause 18. The non-transient, computer-readable storage medium of any of clauses 10-17, wherein the generation of the plurality of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an internal virtual safety bubble during the activation of a landscaping mechanism of the autonomous cutting vehicle, wherein the violation of the internal virtual safety bubble triggers the termination of the landscaping mechanism activation.

[0231] Clause 19. The non-transient, computer-readable storage medium of any of clauses 10 to 18, wherein the execution of the autonomous cutting vehicle operation comprises: detecting a potential violation by the virtual buffer of the plurality of virtual safety bubbles based on a trajectory of the autonomous cutting vehicle; in response to the detection of the potential violation, modifying the configuration of the autonomous cutting vehicle to avoid the violation of the plurality of virtual safety bubbles; and modifying the plurality of virtual safety bubbles around the autonomous cutting vehicle based on the modified configuration of the autonomous cutting vehicle.

[0232] Clause 20. A computer-implemented method comprising: capturing image data from a camera system of a Petition 870250101989, dated 07 / 11 / 2025, page 82 / 122 79 / 87 autonomous cutting vehicle; detect at least one object in an environment around the autonomous cutting vehicle based on image data; generate a virtual buffer for the object, the virtual buffer positioned around the object; generate a plurality of virtual safety bubbles around the autonomous cutting vehicle based on a configuration of the autonomous cutting vehicle; and execute the autonomous operation of the cutting vehicle to perform one or more landscaping actions in the environment, avoiding the violation of the plurality of virtual safety bubbles by the object's virtual buffer.

[0233] Clause 21. The computer-implemented method of clause 20, wherein the image data capture comprises capturing images from stereoscopic pairs of cameras positioned around the autonomous cutting vehicle to capture a 360-degree panoramic view of the environment around the autonomous cutting vehicle.

[0234] Clause 22. The computer-implemented method of any of clauses 20-21, further comprising: generating a spatial representation of the autonomous cutting vehicle based on image data, wherein the spatial representation spatially describes the object in relation to the autonomous cutting vehicle.

[0235] Clause 23. The computer-implemented method of clause 22, further comprising: applying one or more machine learning models to classify an object type from among a plurality of object types for the object based on image data, wherein the virtual buffer for the object is based on the object type classified by one or more machine learning models.

[0236] Clause 24. The computer-implemented method of clause 23, wherein the plurality of object types includes persons and inanimate objects, wherein, in response to classifying the object as a person, the virtual buffer is generated for the object that has a first size, and wherein, in response to classifying the object as an object Petition 870250101989, dated 07 / 11 / 2025, page 83 / 122 80 / 87 inanimate object, the virtual buffer is generated for the object that has a second size smaller than the first size.

[0237] Clause 25. The computer-implemented method of any of clauses 20 to 24, wherein the generation of the plurality of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an outer virtual safety bubble, wherein the violation of the outer virtual safety bubble triggers one or more distance warnings from the autonomous cutting vehicle.

[0238] Clause 26. The computer-implemented method of clause 25, whereby the external virtual security bubble is configured to trigger, as one or more warnings, an audible signal by a loudspeaker or a visual signal by a lighting system.

[0239] Clause 27. The computer-implemented method of any of clauses 25 to 26, whereby the external virtual safety bubble is configured to, in response to the expiration of a timer, terminate the autonomous operation of the autonomous cutting vehicle.

[0240] Clause 28. The computer-implemented method of any of clauses 20-27, wherein the generation of the plurality of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an internal virtual safety bubble during the operation of a landscaping mechanism of the autonomous cutting vehicle, wherein the violation of the internal virtual safety bubble triggers the termination of the landscaping mechanism operation.

[0241] Clause 29. The computer-implemented method of any of clauses 20 to 28, wherein the execution of the autonomous cutting vehicle operation comprises: detecting a potential violation by the virtual buffer of plurality of virtual safety bubbles based on a trajectory of the autonomous cutting vehicle; in response to the detection of the potential violation, modifying the configuration of the autonomous cutting vehicle. Petition 870250101989, dated 07 / 11 / 2025, page 84 / 122 81 / 87 to avoid violating the plurality of virtual safety bubbles; and modify the plurality of virtual safety bubbles around the autonomous cutting vehicle based on the modified configuration of the autonomous cutting vehicle. XIV. ADDITIONAL CONSIDERATIONS

[0242] Throughout this Descriptive Report, multiple instances may implement components, operations, or structures described as a single instance. Although individual operations of one or more methods are illustrated and described as separate operations, one or more of the individual operations may be executed simultaneously, and nothing requires that the operations be executed in the order illustrated. Structures and functionalities presented as separate components in example configurations may be implemented as a combined structure or component. Similarly, structures and functionalities presented as a single component may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of the subject matter discussed herein.

[0243] Certain embodiments are described herein as including logic or a series of components, modules, or mechanisms. Modules may constitute software modules (e.g., code embedded in a machine-readable medium or in a transmission signal) or hardware modules. A hardware module is a tangible unit capable of performing certain operations and may be configured or arranged in a particular manner. Examples of embodiments include one or more computer systems (e.g., a stand-alone computer system, client, or server) or one or more hardware modules of a computer system (e.g., a processor or a group of processors). Petition 870250101989, dated 07 / 11 / 2025, page 85 / 122 82 / 87 can be configured by software (e.g., an application or part of an application) as a hardware module that operates to perform certain operations, as described herein.

[0244] In various embodiments, a hardware module can be implemented mechanically or electronically. For example, a hardware module may comprise dedicated circuits or logic that are permanently configured (e.g., as a purpose-built processor, such as a programmable gate array (FPGA) for landscape environments or an application-specific integrated circuit (ASIC)) to perform certain operations. A hardware module may also comprise programmable logic or circuits (e.g., as those encompassed in a general-purpose processor or other programmable processor) that are temporarily configured by software to perform certain operations. It should be understood that the decision to implement a hardware module mechanically, in dedicated and permanently configured circuits, or in temporarily configured circuits (e.g., configured by software), may be motivated by cost and time considerations.

[0245] Consequently, the term “hardware module” should be understood as encompassing a tangible entity, whether a physically constructed entity, permanently configured (e.g., physically connected) or temporarily configured (e.g., programmed) to operate in a particular manner or to perform certain operations described herein. As used herein, “hardware-implemented module” refers to a hardware module. Considering modalities in which hardware modules are temporarily configured (e.g., programmed), each of the hardware modules does not need to be configured or instantiated at any instance in time. For example, when hardware modules Petition 870250101989, dated 07 / 11 / 2025, page 86 / 122 83 / 87 comprise a general-purpose processor configured using software; the general-purpose processor can be configured as different respective hardware modules at different times. The software can therefore configure a processor, for example, to constitute a specific hardware module at a given time and to constitute a different hardware module at a different time.

[0246] Hardware modules can provide and receive information from other hardware modules. Consequently, the hardware modules described can be considered communicatively coupled. When several such hardware modules exist simultaneously, communication can be achieved through the transmission of signals (e.g., via appropriate circuits and buses) connecting the hardware modules. In embodiments where multiple hardware modules are configured or instantiated at different times, communication between these hardware modules can be achieved, for example, through the storage and retrieval of information in memory structures to which the various hardware modules have access. For example, a hardware module might perform an operation and store the output of that operation in a memory device to which it is communicatively coupled.An additional hardware module can then subsequently access the memory device to retrieve and process the stored output. Hardware modules can also initiate communication with input or output devices and can operate on a resource (e.g., a collection of information).

[0247] The various operations of the exemplary methods described herein may be performed, at least partially, by one or more processors configured temporarily (e.g., by software) or permanently to perform the relevant operations. Whether configured temporarily or permanently, such processors may Petition 870250101989, dated 07 / 11 / 2025, page 87 / 122 84 / 87 constitute processor-implemented modules that operate to execute one or more operations or functions. The modules mentioned here may, in some exemplary embodiments, comprise processor-implemented modules.

[0248] Similarly, the methods described herein may be at least partially implemented by processor. For example, at least some of the operations of a method may be performed by one or more processors or processor-implemented hardware modules. The performance of certain operations may be distributed among one or more processors, not only residing on a single machine, but also deployed across one or more machines, for example, the 700 computer system. In some exemplary embodiments, the processor(s) may be located in a single location (e.g., in a home environment, an office environment, or as a set of servers), while in other embodiments the processors may be distributed across multiple locations.

[0249] The processor(s) may also operate to support the performance of relevant operations in a “cloud computing” environment or as “software as a service” (SaaS). For example, at least some of the operations may be performed by a group of computers (such as machines that include processors), with these operations being accessible via a network (e.g., the Internet) and via one or more appropriate interfaces (e.g., application programming interfaces (APIs)).

[0250] The execution of certain operations can be distributed among one or more processors, not just residing on a single machine, but deployed across multiple machines. It should be noted that when an operation is described as being executed by “one processor,” this should be interpreted as also including the process being executed. Petition 870250101989, dated 07 / 11 / 2025, pages 88 / 122 85 / 87 by more than one processor. In some exemplary embodiments, one or more processors or processor-implemented modules may be located in a single geographic location (e.g., in a home environment, an office environment, or a server array). In other exemplary embodiments, one or more processors or processor-implemented modules may be distributed across multiple geographic locations.

[0251] Some parts of this Descriptive Report are presented in terms of algorithms or symbolic representations of operations on data stored as bits or binary digital signals within machine memory (e.g., computer memory). These algorithms or symbolic representations are examples of techniques used by people with common knowledge in the arts of data processing to convey the essence of their work to other specialists in the field. As used herein, an “algorithm” is a self-consistent sequence of operations or similar processing that leads to a desired result. In this context, algorithms and operations involve the physical manipulation of physical quantities. Typically, but not necessarily, such quantities may take the form of electrical, magnetic, or optical signals capable of being stored, accessed, transferred, combined, compared, or otherwise manipulated by a machine.Sometimes, it is convenient, mainly for reasons of common usage, to refer to such signals using words like "data," "content," "bits," "values," "elements," "symbols," "characters," "terms," ​​"numbers," "numerals," or similar. These words, however, are merely convenient labels and should be associated with appropriate physical quantities.

[0252] Unless otherwise indicated, discussions contained herein that use words such as “processing”, “computing”, “calculating”, “determining”, “presenting”, “displaying” or similar may refer to actions or Petition 870250101989, dated 07 / 11 / 2025, page 89 / 122 86 / 87 processes of a machine (e.g., a computer) that manipulates or transforms data represented as physical quantities (e.g., electronic, magnetic, or optical) within one or more memories (e.g., volatile memory, non-volatile memory, or a combination thereof), registers, or other machine components that receive, store, transmit, or display information.

[0253] As used herein, any reference to “a modality” or “a modality” means that a specific element, resource, structure, or feature described in connection with the modality is included in at least one modality. Occurrences of the expression “in a modality” in various places in the report do not necessarily refer to the same modality.

[0254] Some embodiments may be described using the expressions “coupled” and “connected,” along with their derivatives. It should be understood that these terms are not intended to be synonymous with each other. For example, some embodiments may be described using the term “connected” to indicate that two or more elements are in direct physical or electrical contact with each other. In another example, some embodiments may be described using the term “coupled” to indicate that two or more elements are in direct physical or electrical contact. The term “coupled,” however, may also mean that two or more elements are not in direct contact with each other, but still cooperate or interact with each other. Embodiments are not limited to this context.

[0255] As used herein, the terms “comprises”, “comprising”, “includes”, “including”, “has”, “having” or any other variation thereof, are intended to encompass a non-exclusive inclusion. For example, a process, method, article or device comprising a list of elements is not necessarily limited to only those elements, but may include other elements not expressly listed or inherent to such. Petition 870250101989, dated 07 / 11 / 2025, pp. 90 / 122 87 / 87 process, method, article or apparatus. Furthermore, unless expressly indicated otherwise, “or” refers to an inclusive “or” and not an exclusive “or”. For example, a condition A or B is satisfied by any of the following conditions: A is true (or present) and B is false (or absent), A is false (or absent) and B is true (or present), and A and B are true (or present).

[0256] In addition, the use of “a” or “an” is employed to describe elements and components of the embodiments described herein. This is done merely for convenience and to give a general sense of the invention. This description should be interpreted as including one or at least one, and the singular also includes the plural, unless it is obvious that the contrary is intended.

[0257] Upon reading this disclosure, those skilled in the art will also appreciate additional alternative structural and functional designs for a system and process for providing change assessment in CMC through the principles disclosed herein. Thus, although specific embodiments and applications have been illustrated and described, it should be understood that the embodiments disclosed are not limited to the precise construction and components disclosed herein. Various modifications, alterations, and variations, which will be evident to those skilled in the art, may be made to the arrangement, operation, and details of the method and apparatus disclosed herein, without departing from the principles disclosed herein. Petition 870250101989, dated 07 / 11 / 2025, pp. 91 / 122

Claims

1 / 10 CLAIMS 1. Autonomous Cutting Vehicle, characterized in that it comprises: a camera system comprising cameras positioned around the autonomous cutting vehicle; a cutting platform comprising one or more motorized blades for cutting plants in an environment; and a control system configured to: capture image data from a camera system of an autonomous cutting vehicle; detect an object in an environment around the autonomous cutting vehicle based on the image data; generate a virtual buffer for the object, the virtual buffer positioned around the object; generate virtual safety bubbles around the autonomous cutting vehicle based on a configuration of the autonomous cutting vehicle; and execute, through at least the cutting platform, the autonomous operation of the cutting vehicle to perform one or more landscaping actions in the environment, avoiding the violation of the virtual safety bubbles by the object's virtual buffer.

2. Autonomous Cutting Vehicle, according to Claim 1, characterized in that the control system is further configured to: Petition 870250082523, dated 12 / 09 / 2025, page 10 / 35 2 / 10 generate a spatial representation of the autonomous cutting vehicle based on image data, wherein the spatial representation spatially describes the object in relation to the autonomous cutting vehicle.

3. Autonomous Cutting Vehicle, according to Claim 2, characterized in that the control system is further configured to: apply one or more machine learning models to classify an object type among object types for the object based on image data, wherein the virtual buffer for the object is based on the object type classified by one or more machine learning models.

4. Autonomous Cutting Vehicle, according to Claim 3, characterized in that the object types include people and inanimate objects, wherein, in response to classifying the object as a person, a virtual buffer is generated for the object with a first size, and wherein, in response to classifying the object as an inanimate object, a virtual buffer is generated for the object with a second size smaller than the first size.

5. Autonomous Cutting Vehicle, according to any of Claims 1 to 4, characterized in that the control system, configured to generate virtual safety bubbles based on the configuration of the autonomous cutting vehicle, comprises being configured to: generate an external virtual safety bubble; and Petition 870250082523, dated 12 / 09 / 2025, page 11 / 35 3 / 10 in response to the violation of the external virtual safety bubble, trigger one or more warnings to distance the object from the autonomous cutting vehicle.

6. Autonomous Cutting Vehicle, according to Claim 5, characterized in that the control system, configured to trigger one or more warnings, comprises being configured so that: the external virtual safety bubble is configured to trigger, as one or more warnings, an audible signal via a loudspeaker or a visual signal via a lighting system.

7. Autonomous Cutting Vehicle, according to any one of Claims 5 to 6, characterized in that the control system, configured to generate virtual safety bubbles based on the configuration of the autonomous cutting vehicle, further comprises being configured to: in response to the violation of the external virtual safety bubble, start a timer; and in response to the expiration of the timer, terminate the autonomous operation of the autonomous cutting vehicle.

8. Autonomous Cutting Vehicle, according to any of Claims 1 to 7, characterized in that the control system, configured to generate virtual safety bubbles based on the configuration of the autonomous cutting vehicle, comprises being configured to: generate an internal virtual safety bubble during the activation of the autonomous cutting vehicle's cutting platform; and Petition 870250082523, dated 12 / 09 / 2025, page 12 / 35 4 / 10 in response to the violation of the internal virtual safety bubble, terminate the activation of the cutting platform.

9. Autonomous Cutting Vehicle, according to any one of Claims 1 to 8, characterized in that the control system configured to perform the autonomous operation of the cutting vehicle comprises being configured to: detect a possible violation by the virtual buffer of the virtual safety bubbles based on the trajectory of the autonomous cutting vehicle; in response to the detection of the possible violation, modify the configuration of the autonomous cutting vehicle to avoid the violation of the virtual safety bubbles; and modify the virtual safety bubbles around the autonomous cutting vehicle based on the modified configuration of the autonomous cutting vehicle.

10. Non-Transient Computer-Readable Storage Medium, characterized in that it stores instructions that, when executed by a computer processor, cause the computer processor to perform operations comprising: capturing image data from a camera system of an autonomous mowing vehicle; detecting at least one object in an environment around the autonomous mowing vehicle based on the image data; generating a virtual buffer for the object, the virtual buffer positioned around the object; Petition 870250082523, dated 12 / 09 / 2025, page 13 / 35 5 / 10 generating virtual safety bubbles around the autonomous mowing vehicle based on a configuration of the autonomous mowing vehicle; and executing the autonomous operation of the mowing vehicle to perform one or more landscaping actions in the environment, avoiding the violation of the plurality of virtual safety bubbles by the object's virtual buffer.

11. Computer-Readable Non-Transient Storage Medium, according to Claim 10, characterized in that the image data capture comprises capturing images from stereoscopic camera pairs positioned around the autonomous cutting vehicle to capture a 360-degree panoramic view of the environment surrounding the autonomous cutting vehicle.

12. Non-Transient Computer-Readable Storage Medium, according to any one of Claims 10 to 11, characterized in that the operations further comprise: generating a spatial representation of the autonomous cutting vehicle based on image data, wherein the spatial representation spatially describes the object in relation to the autonomous cutting vehicle.

13. Non-Transient Computer-Readable Storage Medium according to Claim 12, characterized in that the operations further comprise: applying one or more machine learning models to classify an object type among object types for the object based on image data, wherein the virtual buffer for the object is based on the object type classified by one or more machine learning models. Petition 870250082523, dated 12 / 09 / 2025, p. 14 / 35 6 / 10 14. Computer-Readable Non-Transient Storage Medium, according to Claim 13, characterized in that the object types include people and inanimate objects, wherein, in response to classifying the object as a person, a virtual buffer is generated for the object that has a first size, and wherein, in response to classifying the object as an inanimate object, a virtual buffer is generated for the object that has a second size smaller than the first size.

15. Non-Transient Computer-Readable Storage Medium, according to any one of Claims 10 to 14, characterized in that the generation of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an external virtual safety bubble, wherein the violation of the external virtual safety bubble triggers one or more distance warnings from the autonomous cutting vehicle.

16. Non-Transient Computer-Readable Storage Medium, according to Claim 15, characterized in that the external virtual security bubble is configured to trigger, as one or more warnings, an audible signal through a loudspeaker or a visual signal through a lighting system.

17. Non-Transient Computer-Readable Storage Medium, according to any one of Claims 15 to 16, characterized in that the external virtual safety bubble is configured to, in response to the expiration of a timer, terminate the autonomous operation of the autonomous cutting vehicle. Petition 870250082523, dated 12 / 09 / 2025, p. 15 / 35 7 / 10 18. Non-Transient Computer-Readable Storage Medium, according to any one of Claims 10 to 17, characterized in that the generation of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an internal virtual safety bubble during the activation of a landscaping mechanism of the autonomous cutting vehicle, wherein the violation of the internal virtual safety bubble triggers the termination of the landscaping mechanism activation.

19. Computer-Readable Non-Transient Storage Medium, according to any one of Claims 10 to 18, characterized in that the execution of the autonomous cutting vehicle operation comprises: detecting a potential violation by the virtual buffer of virtual safety bubbles based on a trajectory of the autonomous cutting vehicle; responding to the detection of the potential violation, modifying the configuration of the autonomous cutting vehicle to avoid the violation of the virtual safety bubbles; and modifying the virtual safety bubbles around the autonomous cutting vehicle based on the modified configuration of the autonomous cutting vehicle.

20. Computer Implemented Method, characterized in that it comprises: capturing image data from a camera system of an autonomous mowing vehicle; Petition 870250082523, dated 12 / 09 / 2025, page 16 / 35 8 / 10 detecting at least one object in an environment around the autonomous mowing vehicle based on the image data; generating a virtual buffer for the object, the virtual buffer positioned around the object; generating virtual safety bubbles around the autonomous mowing vehicle based on a configuration of the autonomous mowing vehicle; and executing the autonomous operation of the mowing vehicle to perform one or more landscaping actions in the environment, avoiding the violation of the virtual safety bubbles by the object's virtual buffer.

21. Computer-Implemented Method according to Claim 20, characterized in that the image data capture comprises capturing images from stereoscopic camera pairs positioned around the autonomous cutting vehicle to capture a 360-degree panoramic view of the environment surrounding the autonomous cutting vehicle.

22. Computer-Implemented Method, according to any one of Claims 20 to 21, characterized in that it further comprises: generating a spatial representation of the autonomous cutting vehicle based on image data, wherein the spatial representation spatially describes the object in relation to the autonomous cutting vehicle.

23. Computer Implemented Method according to Claim 22, characterized in that it further comprises: applying one or more machine learning models to classify an object type among object types for the object based on image data, wherein the virtual buffer for the object is based on the object type classified by one or more machine learning models.

24. Computer Implemented Method according to Claim 23, characterized in that the object types include people and inanimate objects, wherein, in response to classifying the object as a person, a virtual buffer is generated for the object that has a first size, and wherein, in response to classifying the object as an inanimate object, a virtual buffer is generated for the object that has a second size smaller than the first size.

25. Computer-Implemented Method, according to any one of Claims 20 to 24, characterized in that the generation of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an outer virtual safety bubble, wherein the violation of the outer virtual safety bubble triggers one or more distance warnings from the autonomous cutting vehicle.

26. Computer-Implemented Method according to Claim 25, characterized in that the external virtual safety bubble is configured to trigger, as one or more warnings, an audible signal via a loudspeaker or a visual signal via a light system.

27. Computer Implemented Method, according to any of Claims 25 to 26, characterized in that the external virtual security bubble is configured to, in response to the expiration of a timer, terminate the autonomous operation of the autonomous cutting vehicle.

28. Computer-Implemented Method, according to any one of Claims 20 to 27, characterized in that the generation of virtual safety bubbles based on the configuration of the autonomous cutting vehicle comprises: generating an internal virtual safety bubble during the activation of a landscaping mechanism of the autonomous cutting vehicle, wherein the violation of the internal virtual safety bubble triggers the termination of the landscaping mechanism activation.

29. Computer-Implemented Method, according to any one of Claims 20 to 28, characterized in that the execution of the autonomous cutting vehicle operation comprises: detecting a potential violation by the virtual buffer of virtual safety bubbles based on the trajectory of the autonomous cutting vehicle; in response to the detection of the potential violation, modifying the configuration of the autonomous cutting vehicle to avoid violating the plurality of virtual safety bubbles; and modifying the virtual safety bubbles around the autonomous cutting vehicle based on the modified configuration of the autonomous cutting vehicle. Petition 870250082523, dated 12 / 09 / 2025, p. 19 / 35

Images