Method for controlling the drive of a mobile robot

Abstract

The objective is a drive control method for a mobile robot (1) having multiple wheeled legs (3) with uncontrollable wheels (8) that are more efficient and reliable. This problem is solved by (a) a step of accessing inputs, including targets, provided via a user interface by planning software (111) on computer hardware (11), and a step of determining a plan according to target and environment data maps; (b) a step of operating the hip joints (4), knee joints (6), and wheel motors (9) of all wheeled legs (3) according to the plan by execution software (101) on the electronics (10) of the mobile robot (1); and (c) a step of controlling in real time by the execution software (101) and planning software (111).

Description

Technical Field

[0001] The present invention relates to a driving control method for a mobile robot according to the preamble of claim 1, the use of a neural network for a driving control method of a mobile robot, and a mobile robot.

Background Art

[0002] In US2021276642, a modular robotic vehicle (MRV) is disclosed, which includes a chassis and a body having any shape and dimensions including an enclosed cab, where multiple passengers sit inside the enclosed cab or one passenger sits on a seat without an enclosed cab. The modular chassis of the vehicle further includes a leg array rotatably connected therein, and the leg array includes actuators that cause bending and vertical movement to maintain the MRV stably when moving on various grounds in indoor or outdoor environments. The leg array provides walking and steering capabilities that allow the MRV to cross during navigation mode, and the wheels, by providing differential steering propulsion or braking capabilities, the wheels act like feet when the power is cut off during walking mode and rotate when the power is turned on during driving mode, and the MRV transports passengers and / or cargo.

[0003] US2022097785, CN112874651, CN111846002, CN110861728 and the paper XP011789464 "Trajectory Optimization for Wheeled-Legged Quadrupedal Robots Driving in Challenging Terrain in IEEE Robotics and Automation Letters by Medeiros, Vivian S. et al." are the progress in the field of mobile robots in the past few years that depend on different driving technologies and methods. The latest prior art is XP011789464. However, this document does not describe a sufficiently reliable and efficient method for driving a mobile robot with unsteerable wheels.

[0004] For example, the prerequisites for wheel maneuverability via the ankle joint are often large and complex hardware solutions. The same applies to realizing rotatable connections between the legs and the body, inspired by many animals, such as in the form of ball joints. Furthermore, standard solutions for leg movement result in a limited set of directly induced movements. Moreover, with regard to the operation of robotic vehicles, rapid, smooth, and efficient movement is largely unattainable through standard pre-programmed manual and automatic responses of robotic vehicles to predefined events.

[0005] Most known robotic vehicles also feature either omnidirectional wheels or a steering motor for each wheel. The former adds complexity to the wheel design, and omnidirectional wheels are known to function properly only on flat surfaces with good interaction with the ground. The latter adds additional hardware components, as it requires one more motor per leg to steer the vehicle. The additional steering motors increase the inertia of the legs, negatively impacting performance and cost. Often, wheels are also added to the base. In this case, the robot can only use the wheels on perfectly flat surfaces, and the mass of the wheels is directly added to the body, reducing the available payload.

[0006] To date, mobile robots with wheeled legs whose wheels are uncontrollable have not been able to be controlled in an efficient and reliable manner. [Overview of the project]

[0007] The object of the present invention is to enable a mobile robot to move quickly, smoothly, and efficiently in all directions with minimal hardware and overcome difficult terrain, by generally providing a mobile robot equipped with a large number of uncontrollable wheels, a method for controlling the drive of the mobile robot, and the use of a neural network for the method for controlling the drive of the mobile robot.

[0008] Modifications and minor adaptations of combinations of the features of the present invention are described in the embodiments for carrying out the invention, shown in the drawings, and included in the claims.

[0009] A further understanding of the various aspects of the present invention can be obtained by referring to the following detailed description in conjunction with the relevant drawings, which are briefly described below.

[0010] Note that in embodiments described separately, the same parts are given the same reference numerals or component names, and the disclosures contained in the entire description can be similarly applied to the same parts having the same reference numerals or component names.

[0011] Preferred exemplary embodiments of the subject matter of the present invention are described below with reference to the accompanying drawings. [Brief explanation of the drawing]

[0012] [Figure 1] A perspective view of a mobile robot equipped with a payload interface is shown. [Figure 2] This shows a perspective view of a mobile robot equipped with a single wheeled leg in a step position. [Figure 3] This shows a top view of a mobile robot equipped with a single wheeled leg in a step position. [Modes for carrying out the invention]

[0013] Automated vehicles, such as Mobile Robot 1, are becoming increasingly important not only in the field of transporting goods and people, but also, for example, for exploring unfamiliar terrain. To use such automated vehicles advantageously compared to conventional solutions, they must be fast and efficient, and capable of overcoming obstacles such as stairs, inclined surfaces, wet surfaces, icy surfaces, grass, pebbles, large stones, and other uneven ground, as well as traffic. This presents the challenge of providing hardware that is as simple as possible in terms of operation and manufacturing, as error-free as possible, as light as possible, and enables a physical model of vehicle movement that is as simple as possible. On the other hand, there is a need for high-performance software, computer devices, and electronic equipment for drive control methods that operate automated vehicles, capable of pursuing a given objective while simultaneously reacting in real time to events affecting the vehicle's state. Furthermore, drive control must optimally utilize the characteristics of the hardware, and vice versa. This has been achieved here, as described below.

[0014] The mobile robot 1 presented in this specification is shown exemplarily in Figures 1 to 3.

[0015] Figure 1 shows a mobile robot 1 comprising a torso 2 and four wheeled legs 3. Each wheeled leg 3 includes a hip joint 4, an upper leg 5, a knee joint 6, a lower leg 7, and a wheel 8 with a mounted wheel motor 9. For clarity, in Figure 1, reference numerals corresponding to wheeled legs 3 are shown for only two of the wheeled legs 3. The mobile robot 1 further comprises at least one energy source, at least one sensor, a communication system, and electronics 10. The electronics 10 is operably connected to at least one energy source, at least one sensor, the communication system, and to the hip joints 4, knee joints 6, and wheel motors 9 of all wheeled legs 3. Further equipment and additional wheeled legs 3 are possible.

[0016] The energy source here is a battery, which is operably connected to the electronic device 10. Alternatively, the energy source may be, for example, an internal combustion engine combined with a fuel tank, and an energy converter having electrical energy as output and connected to the electronic device 10, or a combination of such an engine and a battery.

[0017] Here, the torso 2 is formed into a shallow, elongated rectangular prism with recesses and protrusions. On each of the two long, narrow sides of the shallow rectangular prism, one of the wheeled legs 3 is attached within each of the two outer longitudinal quarters of the side. More precisely, the torso 2 is connected to the hip joint 4 of each of the four wheeled legs 3, and the hip joint 4 is then connected to the upper thigh 5 of the corresponding wheeled leg 3. Other shapes of the torso 2 and other arrangements of the wheeled legs 3 are also possible.

[0018] The hip joint 4 of all wheeled legs 3 is here realized by a hip joint component 41 having the shape of a uniform prism. This includes two rotational joints having mutually perpendicular axes, the axes aligned with first and second connectors protruding from the hip joint component 41. Alternatively, the axes can be positioned at different angles to each other, as long as they are not parallel. The first connector protrudes from one of the bases of the prism of the hip joint component 41 toward the center and perpendicular to the base. Its distal end is attached to the hip joint component 41 on the medial surface of the corresponding upper thigh 5, within the uppermost longitudinal quarter of the upper thigh 5, and its proximal end is attached to the hip joint component 41 such that the first connector can rotate with respect to its central longitudinal axis, i.e., its direction of extension. Thus, the rotational joint corresponding to the first connector allows the corresponding wheeled leg 3 to rotate in a plane p_1 substantially parallel to the base of the prism of the hip joint component 41. Here, it should be noted that "approximately parallel" and "approximately perpendicular" are understood to mean that the planes can have angles of up to 15 degrees with respect to truly parallel and truly perpendicular planes, respectively, in the axis of rotation and in any direction. More precisely, it should be noted that different points of the wheeled leg 3 can move on different approximately parallel planes p_1.

[0019] The second connection protrudes from the hip component 41 along a central axis parallel to the base of the prism of the hip component 41. Its proximal end is fixedly attached to the hip component 41, and its distal end is attached to the torso 2 so that the second connection can rotate with respect to its central longitudinal axis, i.e., its direction of extension. Thus, the wheeled leg 3 can rotate in a plane p_2 that is substantially perpendicular to p_1 by the rotational joint corresponding to the second connection. Naturally, the range of this rotation is limited by the torso 2 intersecting this plane. More precisely, it should be noted that different points of the wheeled leg 3 can move in different substantially parallel planes p_2.

[0020] For example, other realizations of the hip joint 4 are possible, such as using a hip joint component 41 of an alternative shape, having no hip joint component 41, having a different arrangement of the rotational joint with respect to relative distance and angle, or having a different arrangement with respect to attachment to the torso 2.

[0021] Here, the knee joint 6 of all wheeled legs 3 is realized as a rotational joint that allows for a change in angle between the corresponding upper leg 5 and the corresponding lower leg 7, within approximately the plane p_1. A possible alternative realization of the knee joint 6 includes a sliding joint, which would allow for a reduction in the distance from the top of the upper leg 5 to the corresponding wheel 8.

[0022] The hip joints 4, knee joints 6, and wheel motors 9 of all wheeled legs 3 are preferably electrically operated. Alternatively, one or more of them may be hydraulic or pneumatic.

[0023] Here, the wheels 8 of each wheeled leg 3 are mounted on the corresponding lower leg 7 such that their axis of rotation lies in a plane substantially parallel to the support of the wheel 8. The wheel motor 9 of the wheel 8 is preferably mounted on the wheel hub of the wheel 8 and is used to accelerate the wheel 8, i.e., to adjust the rotational speed of the wheel 8. In detail, there is no motor for steering each wheel 8; i.e., there is no steering mechanism for the wheel 8, and the wheel 8 is unsteerable. The position of the wheel 8 in the Cartesian x, y, and z directions can be controlled by the joint architecture and three-dimensional active suspension of the wheeled leg 3. In other words, the axis of rotation of the wheel 8 is fixed relative to the lower leg 7. The turning motion of the unsteerable wheel 8 is not provided by a mechanism or can be modified by a mechanism. In this application, the inventors envision a mobile robot 1 having such unsteerable wheels 8 in which optimized control is achieved. This means that further mechanical control and adapted wheels to be controlled do not play a role. The result here is a software-guided solution.

[0024] The body 2 here includes the electronic device 10 and a battery, and the battery is operably connected to the electronic device 10. The electronic device 10 includes a microchip, and the microchip is operably connected to all the wheel motors 9 and the actuators of all the hip joints 4 and all the knee joints 6. These operable connections, as well as the structure of the joints and the wheel motors 9 themselves, are realized such that each rotational joint of the hip joint 4, the knee joint 6, and the wheel motor 9 of each wheeled leg 3 can operate independently of each other and independently of all the hip joints 4, knee joints 6, and wheel motors 9 of all the other wheeled legs 3.

[0025] The two non - parallel rotational joints of the hip joint 4, together with the individual positioning possibilities of the two rotational joints of the hip joint 4 and the knee joint 6, and the individual accelerations of the wheels 8 by the wheel motors 9, form a three - dimensional active suspension system or all the wheeled legs 3.

[0026] The degrees of freedom of movement provided by the two non - parallel rotational joints of the hip joint 4, together with the individual positioning possibilities of all the rotational joints of all the hip joints 4 and all the knee joints 6, and the individual accelerations of all the wheels 8 by the wheel motors 9, are characterized by efficient direction changes, driving, stepping, and switching between driving during stepping, and stabilization. Specifically, this enables omnidirectional movement. For example, stairs can be traversed by stepping, rotating, or a combination of them, and direct movement in a direction not aligned with the wheel alignment can be achieved by stepping. Here, stabilization also includes placing two or more wheels 8 on the ground and balancing, for example, in an upright position where two wheels 8 are in the air.

[0027] More precisely, movement can be achieved by a combination of (a) the concept of individual accelerations of each wheel 8 and (b) the concept of the positioning of the wheeled legs with respect to the surface and the body 2. Note that (b) can also be combined with rotational movement for changing direction and adaptation of the suspension system, for example, with respect to stepping motion.

[0028] For example, changes in direction can be achieved by (a) the concept of utilizing friction and different driving torques at different positions, and (b) the concept of utilizing mass distribution, momentum, and inertia.

[0029] The versatility of the hip joint 4, along with the individual operational capabilities of all joints and all wheel motors 9, allows for a large degree of freedom of movement, for example, by steering each wheel 8 by moving its axis of rotation relative to the plane p_1 without requiring additional joints or motors. Furthermore, the versatility of the hip joint 4 is achieved by two relatively simple and robust joints instead of a fragile and complex solution. Moreover, these embodiments have the advantage of minimal hardware involvement, which means reduced cost and effort and less weight to be carried. The latter fact, due to the reduced weight itself and the reduced computing power required, has less impact on performance considered in the physical model and reduces energy consumption.

[0030] The microchip of the electronic device 10 is also operably connected to at least one sensor and a communication system.

[0031] At least one sensor, here, includes at least a depth sensor and / or camera and / or an inertial measuring unit (IMU) on each side of the torso 2. Optionally, for example, one or more lasers of the wheel motors 9 and / or one or more motor sensors and / or one or more joint position encoders for detecting the state of the joints and / or one or more actuator sensors for the joints and / or one or more mass spectrometers and / or 360-degree cameras on the top surface of the torso 2 and / or a gyroscope and / or position sensor using Global Positioning System (GPS) technology may be incorporated into or attached to the mobile robot 1. All sensors may be incorporated into or attached to one or more of the torso 2 and / or wheeled legs 3.

[0032] The communication system here preferably integrates a wireless transmitter and a wireless receiver, using wireless local area network (WLAN) technology. Other possibilities include next-generation mobile network (NGMN) technologies such as 4G and 5G, Bluetooth low energy (BLE) technology, and wireless personal area network (WPAN) technology. The communication system can also operate with multiple technologies. Since BLE and WPAN typically have shorter ranges than WLAN and NGMN, a combination of long-range and short-range technologies may be advantageous in some cases to cover simultaneous communication to nearby devices, such as remote control, to distant external devices, such as servers. The communication system is preferably attached to or incorporated into the fuselage 2.

[0033] The electronic equipment 10 may include components incorporated into the body 2 and / or components incorporated into one or more of the wheeled legs 3. The electronic equipment 10 may have several microchips instead of one, and in some cases, tasks can be divided among them.

[0034] A preferred application of the mobile robot 1 of the present invention is the transport of articles and people, most preferably articles. As shown in Figure 1, the article transport platform 21 can be attached, for example, to the upper surface of the body 2 of the mobile robot 1 for this purpose. Alternatively, or further, the body 2 may include cavities and / or recesses for transport means. Another preferred application of the mobile robot 1 of the present invention is the collection of large amounts of data by its at least one sensor.

[0035] Figure 2 shows the mobile robot 1 of Figure 1 without the transport platform 21. Furthermore, computer hardware 11 in the form of a laptop 11 is shown. It should be noted that the computer hardware 11 may be, for example, a general personal computer, a server, or one or more microchips incorporated into the mobile robot 1, or a combination of these options. In detail, the computer hardware 11 may also form part of the electronic equipment 10.

[0036] The laptop 11 communicates with the electronic equipment 10 of the mobile robot 1 via the mobile robot 1's communication system and standard WLAN means incorporated into the laptop 11. Alternatively, if the mobile robot 1's communication system uses other technology, suitable means for that communication may be integrated into or connected to the computer hardware 11. If the computer hardware 11 is part of the electronic equipment 10, communication with other parts of the electronic equipment 10 is part of the electronic equipment 10 itself.

[0037] The user interface, which receives user input, processes it if necessary, and transmits it to planning software 111 on computer hardware 11, is here implemented by a graphical application on laptop 11 and connected to electronic equipment 10 via a communication system.

[0038] Examples of alternative user interface implementations include applications on personal mobile devices, touchscreens integrated into mobile robot 1, or remote controls.

[0039] User interfaces can also be implemented, for example, through auditory or tactile means.

[0040] More generally, the user interface is incorporated into the computer hardware 11 or at least temporarily connected to the computer hardware 11. In the latter case, if the computer hardware 11 forms part of the electronic equipment 10, the user interface can be incorporated into the electronic equipment 10 of the mobile robot 1, i.e., in the electronic equipment 10 outside of the computer hardware 11, or can be incorporated into or attached to the mobile robot 1 outside of the electronic equipment 10, or can be incorporated into or constitute such a device that is not a part of the mobile robot 1. If the user interface is incorporated into or constitutes a separate device, and the computer hardware 11 is physically separate from the electronic equipment 10, the connection between the user interface and the computer hardware 11 can be via the electronic equipment 10 of the mobile robot 1.

[0041] The connection between the user interface and the computer hardware 11 does not need to be continuous, but may be any means of transmitting data from the user interface to the computer hardware 11, possibly in time-shifted steps. For example, the user interface may be a console for programming software, which is loaded once into the computer hardware 11. Alternatively, the user interface may belong to any device and / or program that connects to the planning software 111 directly or through additional devices and / or programs via an application programming interface (API) and possibly additional programs and / or computer systems.

[0042] Communication between the computer hardware 11 and the electronic equipment 10 of the mobile robot 1 is used in the drive control method by which the mobile robot 1 is operated. In this method, planning software 111 is installed on the computer hardware 11 and is connected, at least temporarily, to a user interface, i.e., an application running on a laptop 11 that includes the user interface. To operate the mobile robot 1 and / or start a task, the user can set a target via the user interface. This target is then passed to the planning software 111, which receives the target and then determines a plan according to the target and environment data map.

[0043] Alternatively, the planning software 111 may have already received the input, including the objective, via a user interface before it is loaded into the computer hardware 11, and in some cases, it may have already received it while the planning software itself is being programmed. Note that in the latter case, the user interface may be a console for programming the planning software 111. In that case, the input is not received and accessed, but is simply accessed by the planning software 111 during the execution of the method.

[0044] The objectives may be, for example, a destination such as a target location described by GPS data, a compass direction, a mode such as search, speed, or energy saving, a time such as how long to move in a particular direction or the time to move in a particular direction, or a combination thereof. The objectives may also be more general tasks such as keeping moving and not getting stuck, which can already be incorporated via the console as a user interface when programming the planning software 111.

[0045] The data points in an environmental data map can describe, for example, a street plan, height profile, profile of measured data such as temperature, humidity, light, traffic, motion, or raw data from external sensors. More generally, an environmental data map is a set of data points associated with points on a surface or in space, which may be further combined with metadata and / or timestamps. The data points may be fixed, for example in the case of a road map, or variable, for example in a map reflecting repeated temperature measurements at different points, or a combination of both. The selection of points on a surface that contain data points in an environmental data map may also vary. The measured data may be provided by external sensors or by at least one sensor of mobile robot 1. The external sensor may be fixed or, for example, correspond to another mobile robot.

[0046] The plan is a set of instructions for the electronic equipment 10 of the mobile robot 1. These instructions include commands for the individual actions of each joint and each wheel motor 9, i.e., their movements. The commands specify, for example, the type and intensity of the action. The instructions may include, for example, commands for immediate execution, a list of commands to be executed sequentially, or commands linked to some trigger, or a combination thereof. The trigger may be, for example, a point in time, relative time, or geographical location. In the case of commands to be executed sequentially and / or a list of commands linked to some trigger, the execution of the instructions by the electronic equipment 10 of the mobile robot 1 according to the plan allows the mobile robot 1 to follow a specific trajectory.

[0047] The planning software 111 determines a set of instructions that constitute the plan using a learning method that employs the concepts of changing direction and overcoming obstacles by adjusting the position of joints, and changing direction by adjusting the acceleration of wheels. More generally, reinforcement learning, learning, optimal control, and optimization methods can be used, and a motion library can be used. The optimization goals and constraints for determining the plan are given by user-set goals, data from an environmental data map, structural details of the mobile robot 1, and optionally additional available data. The latter includes information on possible payloads provided by sensors, or by the user via a user interface and / or by data from a digital twin. Computationally intensive tasks are most preferably performed here with the involvement of at least one neural network 112 schematically shown in Figure 2. The at least one neural network 112 may be part of the computer hardware 11. Alternatively, the at least one neural network 112 may be implemented in an external device such as a server or server complex that can communicate with the computer hardware 11, for example, via a WLAN. In the case of multiple neural networks, they may be implemented on the same hardware or distributed across different hardware, including computer hardware 11. If the environmental data map is variable, the plan decision is based either on the state of the environmental data map at a given point in time or on the average over a given period.

[0048] After the plan is decided, the planning software 111 starts transmitting the plan to the mobile robot 1's electronic equipment 10 via the mobile robot 1's communication system and computer hardware 11.

[0049] The determination of the plan and its transmission to the electronic equipment 10 of the mobile robot 1 are preferably initiated as soon as the necessary data becomes available.

[0050] It should be noted that user input via the user interface may include more information than just the goals. In that case, this information may also be transmitted to the planning software 111 and processed by the planning software 111, or this information and / or the goals may be transmitted to the electronic device 10 and processed by the electronic device 10, or a combination of these options may be applied.

[0051] After the plan is determined and transmitted to the electronic equipment 10 of the mobile robot 1, the execution software 101 installed in the electronic equipment 10 of the mobile robot 1 receives the plan and subsequently begins executing the instructions included in the plan.

[0052] Simultaneously, or before or after, for example, after a given period, once initiated, the execution software 101 begins orchestrating the execution of measurements using at least one sensor. This is done according to predefined settings of the execution software 101 and / or the electronic equipment 10, and / or according to instructions received from the computer hardware 11. The generated measurement data is transmitted from at least one sensor to the electronic equipment 10 via its operable connection to at least one sensor, or directly from at least one sensor to the computer hardware 11 by a communication device. If the measurement data is transmitted to the electronic equipment 10, the measurement data is processed there first, or transmitted directly to the computer hardware 11 via a communication system. There may be additional measurement data that is processed only by the electronic equipment 10 and not transmitted to the computer hardware 11.

[0053] Planning software 111 on computer hardware 11 receives measurement data and then, based on the measurement data, decides on adjustments to the plan, possibly based on additional data such as, for example, the updated state of the plan and / or variable environmental data maps, and / or the data used to determine the plan and / or measurement data provided by external sensors. The decision to adjust is achieved by an inference process. This can use a learning method that employs the concepts of changing direction and overcoming obstacles by adjusting the position of joints and changing direction by adjusting the acceleration of the wheels 8. This learning method may be the same as the learning method employed for determining the plan. Computationally intensive tasks are preferably outsourced to at least one neural network 112, as in the case of determining the plan. If both determining the plan and determining adjustments to the plan involve at least one neural network 112, these may be the same neural network 112 or different neural networks 112. The adjustments may be a set of instructions for the electronic equipment 10 of the mobile robot 1, each coupled with instructions on which part of the plan should be replaced or executed in parallel. Alternatively, the adjustment may be, for example, a replacement of the plan, or a completely updated version of the plan that is performed in parallel with or after the plan. Adjustments to the plan allow for responses to environmental occurrences that are not captured or were not captured in the environmental data map, such as previously unknown configurations of surface areas that the mobile robot 1 must pass through, and / or variable environmental occurrences such as traffic or humidity. Furthermore, adjustments to the plan can also be used to smooth transitions defined by the plan, for example, between phases of motion when a sequence of commands from the motion library is inserted, or between synthesized motion sequences from the motion library.

[0054] Once the decision is made, adjustments to the plan are initiated by the planning software 111 and transmitted to the mobile robot's electronic equipment 10 via the mobile robot's communication system and computer hardware 11.

[0055] Next, the execution software 101 receives the adaptation to the plan and adjusts the operation of the joint and wheel motors 9 accordingly.

[0056] It should be noted that the steps from collecting measurement data to adjusting the operation according to the plan adjustments may be performed under real-time control and may be repeated periodically or continuously.

[0057] The transmission, calculation, and execution of the data described above are preferably initiated as soon as the necessary data becomes available.

[0058] If the computer hardware 11 is part of the electronic device 10, round-trip data transfer between the planning software 111 and the execution software 101 may be performed by hardware connections within the electronic device 10. If the computer hardware 11 is part of the electronic device 10, the planning software 111 and the execution software 101 may further be implemented in a single software solution, i.e., in a common software system or within a single program.

[0059] If the planning software 111 and the execution software 101 are implemented in a single software solution, data transfer is determined by this single software solution, and the individual steps of the method can be merged. For example, the decision of the plan in step a and the decision of adjustment to the plan in step c can be performed using the same algorithm that performs a continuous reasoning process. This algorithm first takes target and environment data maps as input and calculates and outputs instantaneous commands for joint positions and motion states from them. Meanwhile, the electronic equipment 10 has already collected measurement data from at least one sensor of the mobile robot 1 and provides the measurement data to the algorithm. From the measurement data, target, environment map which may be updated, and possibly additional data, the algorithm further calculates and outputs instantaneous commands for joint positions and motion states. In this case, the determination and execution of commands can be achieved continuously and almost instantaneously. The reasoning process performed by the algorithm can be executed several times per second, and the commands are recalculated each time. This algorithm manages, for example, a mobile robot 1 to perform a specific task by driving, stepping, sliding one or more of its wheeled legs 3, adjusting the position of its joints for maneuvering, and potentially doing all of this simultaneously and without further human guidance.

[0060] In the case of computer hardware 11 including external components such as a server and internal components of the mobile robot 1 such as a microchip, the planning software 111 may be divided into several subroutines. For example, a subroutine that determines the plan may be executed on an external server, and a subroutine that determines adjustments to the plan may be executed on an internal microchip.

[0061] Even after the initial input, which includes the target, there may be further input from the user via the user interface. This could include replacing the target. For example, mobile robot 1 may receive input via remote control.

[0062] Measurement data, plans, adjustments to plans, specific decisions made during plan determination or adjustments to plans, and possibly other data may be collected, possibly processed, and used for future plan determination and plan adjustment. If there are multiple mobile robots 1 that can provide such data, the multiple mobile robots 1 may share such data among themselves.

[0063] Figure 3 shows a top view of the mobile robot 1 in Figures 1 and 2. In detail, the position of the hip joint 4 of the wheeled leg 3 in the step position is shown in comparison to the position of the hip joint 4 of the wheeled leg 3 when it is not in the step position.

[0064] As shown in Figures 1 to 3, the torso 2 of the mobile robot 1 includes a shell that houses electronic equipment 10 and optionally fragile hardware components and protects them from external influences. A cavity may exist in the shell or in the torso 2 beneath the shell for transporting items. Preferably, the shell has an aerodynamic shape.

[0065] Here, the upper leg 5 of each wheeled leg 3 is cylindrical and has an end that increases in thickness towards the hip joint 4. The lower leg 7 of each wheeled leg 3 is S-shaped in the longitudinal direction and has a rectangular cross-section. Alternative configurations include upper leg 5 and lower leg 7 that have the same shape but may be of different sizes.

[0066] In this example, the mobile robot 1 is equipped with a light source at the front of its torso 2.

[0067] The mobile robot 1 presented herein can move in all directions on any terrain without using steerable wheels or wheel steering mechanisms. Its minimal hardware design enables efficient traversal of flat and uneven terrain, versatility in overcoming obstacles such as steps and stairs, and the ability to carry sensors and objects in any indoor and outdoor space. In detail, the articulated architecture and three-dimensional active suspension of the wheeled legs 3 enable a high degree of freedom of movement. They allow control of the position of the wheels 8 in the Cartesian x, y, and z directions.

[0068] The omnidirectional movement of the mobile robot 1 is achieved by deploying an algorithm that adjusts the relative positions of the joints and the speed of the wheels 8, and by utilizing one or more of the concepts of maneuvering, namely, the concept of overcoming friction constraints by adapting ground reaction forces by simultaneously positioning the center of gravity and adapting the wheel speed, the concept of hybrid mobility by simultaneously performing steps and rotations, and the concept of changing direction by adapting the suspension system, namely rotating the robot's torso around the yaw.

[0069] The learning algorithm utilizes the hardware design to automatically identify appropriate maneuvering concepts, real-time control enhances performance against the variability of the physical world, and narrows the gap between simulation and reality when training is performed in addition to or instead of with the actual mobile robot 1.

[0070] This makes it possible to seamlessly track velocity commands that change rapidly in any direction. Other systems, on the other hand, require switching from feet to wheels or rotating steering motors to satisfy the system's non-holonomic constraints. This also allows the user to provide arbitrary input regarding how and when to transition from one steering strategy to another.

[0071] (b) The approach presented herein, which uses a neural network to achieve (a) autonomous motion using uncontrollable wheels while overcoming difficult terrain such as steps and stairs, is the only approach to achieve omnidirectional motion of the mobile robot 1 in real time, i.e., at a high update rate of more than 5 Hz.

[0072] The learning method achieves the (real-time) behavior of an uncontrollable platform. These behaviors include controllable behavior across arbitrary terrain. Furthermore, the solution presented herein is a minimal solution for achieving efficient movement on flat and challenging terrain. Other known methods cannot achieve this and are computationally slow, requiring the deployment of a neural network.

[0073] The motions achieved include, for example, (i) overcoming friction constraints by simultaneously adapting the center of gravity and wheel speed to adapt ground reaction forces, (ii) hybrid mobility by simultaneously performing steps and rotations, and (iii) adapting the suspension system, i.e., changing direction by rotating the robot's torso around the yaw.

[0074] These movements must (i) be observed without heuristic engineering, (ii) without data from other animals since such forms do not exist in nature, and (iii) the movements must be self-discovered by a learning algorithm.

[0075] Unlike other proposed systems, our neural network can simultaneously manage multiple functions, including tracking robot balance, maneuvering, stepping, obstacle overcoming, and speed commands.

[0076] Regarding the use of mobile robot 1 for transport, the ability of the drive control method, which includes including its planned payload from the outset and then, during operation, responding to possible movements of the goods or people being transported by real-time control, ensures robust and efficient execution of the task. Furthermore, the minimal hardware solution offers the advantages of not only fewer manufacturing steps but also lower energy consumption and lower net weight. This allows for larger payloads on the one hand, and on the other hand, the movement of mobile robot 1 enables additional sensors to ensure the safety of itself, its payload, and other traffic components.

[0077] Please note that the explanations given in this section may also apply to one or more features that exhibit characteristics different from those described in this section. [Explanation of symbols]

[0078] 1. Mobile robot 2. Torso 21. Transport platform 22.Light source 3. Wheeled legs 4. Hip joint 41. Hip joint components 5. Upper leg 6. Knee joint 7. Lower leg 8. (Uncontrollable) wheels 9. Wheel motor 10.Electronic equipment 101. Executable Software 11. Computer Hardware 111. Planning Software 112. Neural Networks

Claims

1. A method for controlling the drive of a mobile robot (1) comprising a torso (2) and at least four wheeled legs (3), wherein each wheeled leg (3) comprises a hip joint (4), an upper leg (5), a knee joint (6), a lower leg (7), and an uncontrollable wheel (8) having an attached wheel motor (9), the mobile robot (1) comprising at least one energy source, at least one sensor, a communication system, and electronic equipment (10), the electronic equipment (10) being connected to each of the at least one energy source, the at least one sensor, and the communication system, The hip joints (4), knee joints (6), and wheel motors (9) of all wheeled legs (3) are operably connected, and the hip joints (4) of all wheeled legs (3) include two rotary joints having axes that are not parallel to each other, and each rotary joint of the hip joints (4) of all wheeled legs (3), the knee joints (6), and the wheel motors (9) are operable by the electronic equipment (10) independently of each other and independently of all other hip joints (4), knee joints (6), and wheel motors (9) of all other wheeled legs (3), and the method is, a. By planning software (111) on computer hardware (11) which forms part of the electronic device (10) or is physically separated from the electronic device (10), - Steps provided via a user interface to access inputs including the goal, and - A step of determining a plan in accordance with the aforementioned objectives and environmental data map, b. The execution software (101) on the electronic device (10) of the mobile robot (1) - A step of operating the hip joint (4), knee joint (6), and wheel motor (9) of all wheeled legs (3) in accordance with the above plan, c. A real-time control step, - The execution software (101) The measurement is performed using at least one of the aforementioned sensors, - The planning software (111) To determine adjustments to the plan based on the aforementioned measurement data, - The execution software (101) executes according to the plan and the adjustments received to the plan, To operate the hip joint (4), knee joint (6), and wheel motor (9) of all wheeled legs (3) The step of real-time control Includes, The determination of the plan in step a and / or the determination of the adjustment of the plan in step c are carried out using a learning method that employs the concept of changing direction by adjusting the position of a joint and the concept of changing direction by adjusting the acceleration of the uncontrollable wheel (8), the learning method being carried out with the involvement of at least one neural network (112). The aforementioned method.

2. The method according to claim 1, wherein the determination of the plan in step a and the determination of the adjustment of the plan in step c are carried out using a learning method that employs the concept of changing direction by adjusting the position of the joint and the concept of changing direction by adjusting the acceleration of the uncontrollable wheel (8), the learning method being carried out with the involvement of at least one neural network (112).

3. The method according to one of the prior claims, wherein the planning software (111) and the execution software (101) are implemented in a single software solution.

4. The method according to claim 3, wherein one or more individual steps of the method are merged into one or more combined steps.

5. The method according to one of the prior claims, wherein the execution software (101) orchestrates the execution of the measurement using the at least one sensor.

6. The method according to one of the prior claims, wherein in step c, the planning software (111) determines adjustments to the plan based on the measurement data, which is achieved using an inference process.

7. The use of at least one neural network (112) for implementing a drive control method for a mobile robot (1), wherein the at least one neural network (112) is integrated into computer hardware (11), and a learning algorithm executed on the at least one neural network (112) is employed by planning software (111) executed on the computer hardware (11) and / or execution software (101) executed on the electronic equipment (10) of the mobile robot (1), to independently optimize the operation of the speed of an uncontrollable wheel (8) and the positions of the hip (4) and knee joints (6) of the numerous wheeled legs (3) of the mobile robot (1) with respect to a predefined target.

8. a. By the planning software (111) on the computer hardware (11) which forms part of the electronic device (10) or is physically separated from the electronic device (10), - Detailed steps provided via the user interface to access inputs including the goal, and - Detailed steps for determining the plan in accordance with the aforementioned objectives and environmental data map, b. The execution software (101) on the electronic device (10) of the mobile robot (1) - Detailed steps of operating the hip joint (4), knee joint (6), and wheel motor (9) of all wheeled legs (3) in accordance with the above plan, c. Detailed steps of real-time control, - The execution software (101) The measurement is performed using at least one of the aforementioned sensors, - The planning software (111) To determine adjustments to the plan based on the aforementioned measurement data, - The execution software (101) The detailed steps of the real-time control are to operate the hip joint (4), knee joint (6), and wheel motor (9) of all wheeled legs (3) in accordance with the plan and the adjustments received to the plan, The determination of the plan in step a and / or the determination of the adjustment of the plan in step c are carried out using a learning method that employs the concept of changing direction by adjusting the position of the joint and the concept of changing direction by adjusting the acceleration of the uncontrollable wheel (8), the learning method being carried out with the involvement of at least one neural network (112). Use of at least one neural network (112) as described in claim 7.

9. A mobile robot (1) operated by a drive control method according to one of claims 1 to 6 and / or by the use of at least one neural network (112) according to one of claims 7 to 8, wherein the mobile robot (1) comprises a torso (2) and at least four wheeled legs (3), each wheeled leg (3) comprising a hip joint (4), an upper leg (5), a knee joint (6), a lower leg (7), and an uncontrollable wheel (8) with an attached wheel motor (9), the mobile robot (1) comprises at least one energy source, at least one sensor, a communication system, and electronic equipment (10), the electronic equipment (10) being operably connected to each of the at least one energy source, the at least one sensor, and the communication system, and being operably connected to the hip joint (4), the knee joint (6), and the wheel motor (9) of all wheeled legs (3), The hip joint (4) of all wheeled legs (3) is provided with two rotational joints having axes that are not parallel to each other, Each rotational joint of the hip joint (4), the knee joint (6), and the wheel motor (9) of all wheeled legs (3) are to be able to operate independently of each other and independently of all other hip joints (4), knee joints (6), and wheel motors (9) of all other wheeled legs (3) by the electronic equipment (10). The mobile robot (1) is characterized by the following.

10. The mobile robot (1) according to claim 9, wherein the hip joints (4) of all wheeled legs (3) are realized by a hip joint component (41) in the form of a uniform rectangular prism having first and second connecting portions, the first connecting portion projecting perpendicularly to the bottom from one of the bottoms of the hip joint component (41) toward the center and fixed to one end which is fixed to the elongated side of the upper thigh (5) of the wheeled leg (3) and is rotatably mounted to the hip joint component (41) opposite the longitudinal axis of the center of the hip joint component (41), and the second connecting portion projecting from the hip joint component (41) toward the hip joint component (41) toward one end which is fixed to the hip joint component (41) and is rotatably mounted to the torso (2).

11. The mobile robot (1) according to one of claims 9 to 10, wherein the hip joint (4), knee joint (6), and wheel motor (9) of all wheeled legs (3) are operated by electricity, hydraulics, pneumatics, or a combination of these options.

12. The mobile robot (1) according to one of claims 9 to 11, wherein the at least one sensor includes at least one depth sensor and / or at least one camera and / or at least one inertial measuring unit (IMU).

13. The mobile robot (1) according to claim 12, wherein the at least one sensor includes one or more lasers and / or one or more motor sensors for a wheel motor (9), and / or one or more joint position encoders and / or one or more actuator sensors for a joint, and / or one or more mass spectrometers.

Images