Cluster robot platform and system based on implicit cooperation and multi-dimensional perception

Abstract

The invention relates to a cluster robot platform based on implicit cooperation and multi-dimensional perception, and aims to improve the performance of a cluster robot in the aspects of measurement precision, perception capability and real-time performance. The platform is composed of a cluster robot hardware system and a software system. The hardware system comprises a bottom layer motion system, an optical cooperation beacon system and a main control board. A software system realizes cooperation and efficient behaviors among cluster robots through a multi-neighbor recognition and identity matching mechanism, a cooperative positioning mechanism, a relative speed direction measurement and calculation mechanism and a realistic differentiation information perception mechanism. The robot can accurately identify neighbors, perform calculation and adjust behaviors in real time so as to adapt to complex cluster tasks. The cluster robot platform supports verification and adaptation of various cluster algorithms, has high-precision perception, excellent real-time performance and wide algorithm adaptation capacity, and provides a solid foundation for research, development and application of the cluster robot technology.

Description

Technical Field

[0001] The present invention is mainly applied to the field of swarm robotics, especially a robot swarm behavior control and perception system based on implicit cooperation and multi-dimensional perception. Background Art

[0002] Swarm robotics technology originated from the imitation of the collective behaviors of organisms in nature, such as bird flocks, fish schools, and ant colonies. These groups form complex collective behaviors through local perception and mutual interaction, usually showing higher efficiency and flexibility than individual organisms. Swarm robots aim to achieve characteristics such as decentralization, local perception, and self-organization by imitating these natural behaviors, and are widely used in various tasks, such as intelligent transportation, automated logistics, and environmental monitoring. In recent years, significant progress has been made in the research of swarm algorithms and swarm robotics technology, covering multiple aspects such as theoretical modeling, algorithm design, simulation, and experimental verification. However, with the rapid development of theoretical models, existing swarm robot platforms face challenges in actual deployment and verification, especially in engineering issues such as robot perception, communication, control, and execution performance. Current swarm robot platforms often rely on digital simulation for verification, but this ignores the hardware limitations and environmental complexity in actual applications.

[0003] Therefore, it is particularly important to establish a general swarm robot platform that can adapt to multiple swarm algorithms. Such a platform can not only promote the verification and optimization of swarm algorithms, but also provide a standardized experimental environment for researchers, thus accelerating the practical application of swarm intelligence. Through this platform, researchers can verify multiple swarm algorithms in different task scenarios, evaluate their effectiveness and adaptability, and provide strong support for the development of swarm intelligence. Summary of the Invention

[0004] To solve the above technical problems, the purpose of the present invention is to provide a vision-based air-space integrated cross-domain collaborative perception technology, system, and device, which can combine the air and space domains to achieve the tracking and positioning of multiple non-cooperative targets far from the ground station on the sea surface.

[0005] The present invention proposes a swarm robot platform based on implicit cooperation and multi-dimensional perception, aiming to support the verification and adaptation of multiple swarm algorithms through advanced hardware and software designs. The summary of the invention covers the design and implementation of the swarm robot hardware system and software system, and detailed functional tests and experimental verifications are carried out.

[0006] First is the design of the swarm robot hardware system. The design focus of the hardware system is to achieve high-precision perception and stable movement, and provide a reliable experimental platform through reasonable system integration. The main components include a motion chassis, an optical cooperation beacon system, the integration of key hardware modules, etc.

[0007] Kinetic chassis design: To ensure the stability and precision of the robot in multi-modal motion, a two-wheel differential motion system is selected in this invention. Two stepper motors drive the left and right wheels, and cooperate with the front and rear omnidirectional wheels to form a tower structure. The stepper motor can provide high-precision rotational speed control, and has a fast response speed, adapting to low-speed and high-precision motion, meeting the requirements of the swarm algorithm for the robot's motion. Compared with the brushed DC motor, the stepper motor will not have a low-level dead zone and has better open-loop control performance, enabling the robot to converge towards the target direction more precisely.

[0008] The kinematic model is based on two-wheel differential design and considers two assumptions: one is that the two driving wheels roll along the center line of the robot without generating side slip or sliding friction; the other is that the two driving wheels can be independently accelerated or decelerated to enable the robot to perform rotational or linear motion. This design simplifies the motion control and enhances the measurement accuracy of the relative speed direction.

[0009] Cooperative beacon system design: The optical cooperative beacon system consists of multiple groups of LED beacons, which are used to characterize the morphological features of the robot and achieve precise neighbor perception. The system is divided into two parts: the positioning rod and the heading rod. The positioning rod is equipped with infrared LEDs and blue LEDs, which are responsible for providing the positioning and direction information of the robot. The LED light source on the heading rod is used to determine the heading angle of the robot and assist neighbors in calculating the relative heading. The layout and emission angle of the LEDs are precisely designed to ensure effective signal transmission at different angles.

[0010] This optical beacon system enables the robot to emit optical signals evenly in all directions within a range of 2 meters, enhancing the perception accuracy. The system also has strong anti-interference ability. By reducing the optical path interference between LED groups, it ensures the perception accuracy in complex environments.

[0011] Integration of key hardware modules: The swarm robot uses Jetson Nano as the main control board, equipped with a Sony IMX219-160IR camera, supporting multi-view visual perception. The main control board realizes remote operation and data transmission through a wireless module. The system integrates various sensors, drive modules, and control boards to ensure that the robot can quickly respond to environmental changes and execute tasks.

[0012] Secondly, it is the design of the software system for the photosensitive swarm robot. The software system is the core part of the technical solution of this invention, responsible for processing and making decisions on the perception information of the photosensitive robot. The system supports functions such as multi-neighbor recognition, cooperative positioning, relative speed direction measurement, and phenomenological difference information perception.

[0013] Multi - neighbor Recognition and Identity Matching Mechanism: Through visual sensors, the photosensitive robot can obtain neighbor information in real - time and perform identity matching. Image processing uses HSV color space conversion and contour extraction algorithms to accurately obtain the geometric center of the light spot. By pairing blue light spots, neighbor recognition is achieved. To improve recognition efficiency, the algorithm adopts the minimum bounding rectangle (Body - ROI) method based on multi - neighbor matching and calculates the intersection - over - union (IoU) between neighbors to complete identity matching.

[0014] Cooperative Localization Mechanism: To ensure precise cooperation among swarm robots, the present invention designs a simplified cooperative localization method. By identifying the LED signals on the positioning rods and heading rods of neighbors and combining geometric relationships, the photosensitive robot can calculate the relative distance and orientation of neighbors. This localization algorithm depends on the inherent parameters of the robot and the geometric features of neighbors, which can reduce hardware complexity and improve localization accuracy.

[0015] Relative Velocity Direction Measurement Mechanism: For the swarm algorithm based on velocity consistency, the platform adopts an improved relative velocity direction measurement mechanism. The photosensitive robot calculates the relative velocity direction of neighbors by distinguishing the optical features of the forward and backward heading rods of neighbors. This method reduces the discretized velocity direction angle and improves the accuracy and continuity of velocity measurement.

[0016] Phenomenological Differential Information Sensing Mechanism: In multi - type swarm algorithms, the state information between individuals is often transmitted through the LED flashing frequency. The photosensitive robot can judge the behavior state of neighbors by checking the flashing frequency of neighbors and recording its changes, which helps the swarm algorithm to achieve coordination and optimization.

[0017] Finally, there are swarm algorithm verification and experiments. The robot platform provided by the present invention is used to verify three types of swarm intelligence algorithms: swarm algorithms based on inter - neighbor distance, velocity consistency, and phenomenological differences. Experiments show that the platform can accurately perceive and execute various swarm algorithms, demonstrating excellent performance in multi - neighbor cooperation, dynamic adjustment, and stability. In the swarm algorithm based on inter - neighbor distance, the robots can quickly converge to the desired distance and maintain stability for a long time. In the velocity consistency algorithm, the robots can quickly adjust their respective velocity directions to ensure the coordination of swarm behavior. In the phenomenological difference algorithm, the photosensitive robot can identify the states of neighbors through optical beacons and make adaptive adjustments. Brief Description of the Drawings

[0018] Figure 1 is the integrated effect diagram of the hardware architecture of the photosensitive robot of the present invention; Figure 2 is the conceptual design diagram of the photosensitive robot of the present invention; Figure 3 is the software implementation framework of the photosensitive swarm robot of the present invention. Detailed implementation manners

[0019] The following further elaborates on the present invention in detail in conjunction with the accompanying drawings and specific embodiments. For the step numbers in the following embodiments, they are only set for the convenience of elaboration and explanation, and no limitation is imposed on the order between steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0020] The present invention provides a swarm robot platform based on implicit collaboration and multi-dimensional perception, aiming to support the verification and adaptation of various swarm algorithms and improve the performance of swarm robots in terms of measurement accuracy, perception ability, and real-time performance through innovative designs of the hardware system and software system. The following details the specific implementation manners of the present invention.

[0021] S1. Design and implementation of the swarm robot hardware system. The hardware system design of the swarm robot platform of the present invention takes into account motion accuracy, perception ability, and reliability to ensure that the robots can cooperate efficiently in swarm tasks.

[0022] Specifically, the motion system adopts a two-wheel differential design. Each robot is equipped with two stepper motors to drive the left and right wheels, and stable balance is achieved through two universal wheels at the front and rear. Each wheel is driven by a stepper motor, which has high-precision speed control ability and can accurately respond to low-speed control requirements, especially suitable for scenarios with high-precision requirements in swarm algorithms. Compared with traditional brushed DC motors, stepper motors have higher response accuracy at low speeds and are not prone to low-level dead zones, so they can better adapt to the precise control requirements of swarm algorithms. The overall design is as Figure 1 shown.

[0023] The design of the optical cooperative beacon system is as Figure 2 shown. To achieve high-precision neighbor recognition and perception, the present invention adopts an optical cooperative beacon system. The system consists of multiple LED beacons, which are distributed on the positioning rod and the heading rod. The specific design is as follows: Positioning rod: This part includes three groups of LED beacons with different heights. The red LED and the blue LED each have different light-emitting angles. The light-emitting angle of the red LED is 150°, which is used to provide optical signals to assist the robot in determining its own position and perceiving the distance to neighbors. The light-emitting angle of the blue LED is 125°, which is used to help the robot obtain the relative position and orientation of neighbors.

[0024] Heading rod: Each robot is equipped with a forward heading rod and a rearward heading rod to transmit the robot's heading information and help neighbors calculate the heading angle relative to itself. The infrared LEDs on the heading rod can provide accurate directional information, enabling the accurate identification of the robot's posture.

[0025] The optical beacon system uses LEDs of different colors (such as infrared LEDs and blue LEDs) to ensure effective signal differentiation, and utilizes optical image processing technology for signal reception and analysis. In this way, the robot can perform precise identity matching and positioning based on the optical signals of adjacent neighbors.

[0026] Camera and perception system. The swarm robot uses a Sony IMX219-160IR camera with a 160° viewing angle, which can capture images of the surrounding environment in real time. The camera does not integrate an infrared filter and can receive the optical signals emitted by the LED beacons. The frame rate of image acquisition is 60 frames per second, ensuring real-time perception and dynamic tracking. Through the camera system, the robot can obtain the spot information of its neighbors, which is used to further analyze the relative positions, velocity directions, and other key information of the neighbors.

[0027] Main control board and communication module. Each robot is equipped with a Jetson Nano as the main control board and uses an Intel8265 AC module to provide stable wireless communication functions. The main control board realizes remote upload and backup of code through management software, and executes swarm algorithms and perception tasks through its embedded computing capabilities. Jetson Nano supports acceleration libraries such as deep learning and computer vision, providing powerful AI capabilities for the robot platform.

[0028] S2. Design and implementation of the swarm robot software system. The software system design of the swarm robot platform of the present invention includes multiple core modules, which are responsible for controlling the swarm behavior of the robot, processing perception information, controlling movement, and executing swarm algorithms. The overall software design framework is as Figure 3 shown.

[0029] First, the multi-neighbor recognition and identity matching mechanism. To ensure that the robot can recognize and distinguish its neighbors, the software system uses a multi-neighbor recognition and identity matching mechanism. The specific implementation steps are as follows: After the robot captures image data through the camera, it first performs image preprocessing, converts it to the HSV color space, and removes noise through techniques such as band-pass filtering and Gaussian blurring.

[0030] In the preprocessed image, the spots in the image are recognized through a contour extraction algorithm, and the geometric center coordinates of each spot are extracted through a polygon approximation algorithm.

[0031] By pairing the blue spots in the image, the identities of the neighbors are determined, and the relative distances and azimuth angles between the neighbors are calculated based on prior knowledge (for example, the actual distance between the blue LEDs of the positioning rods). This mechanism can accurately perform neighbor identity matching and avoid misidentification caused by spot overlap or signal interference.

[0032] Secondly, the cooperative positioning mechanism is one of the key technologies in the present invention. A simplified image feature point-based positioning method is adopted, which reduces the hardware complexity and improves the positioning accuracy. The specific steps are as follows: The robot observes the LED light spots on the positioning rod and the heading rod of its neighbors through a camera, and combines geometric relationships. By using the known LED positions and camera viewing angle parameters, it calculates the relative positions and orientations of the neighbors.

[0033] Triangulation is carried out using geometric relationships to calculate the relative distances and azimuth angles of neighboring robots, and then determine their positions in the cluster. This cooperative positioning mechanism can efficiently calculate the positions of neighbors without occlusion and provide accurate inputs for the clustering algorithm.

[0034] Furthermore, the relative velocity direction measurement mechanism. To ensure the coordination of swarm robots during swarm behavior execution, the software system calculates the velocity directions of neighboring robots through the relative velocity direction measurement mechanism. The specific process is as follows: The robot distinguishes the optical features of the forward and rearward heading rods of neighboring robots through image processing technology and determines the velocity directions of the neighbors based on visual information.

[0035] By calculating the ratio of the pixel distance of the light spot to the actual distance, the robot can accurately measure the velocity direction angle of the neighbor. This mechanism can provide accurate relative velocity direction measurements and help the robot quickly adjust its movement direction in the swarm.

[0036] Finally, the phenomenological differential information perception mechanism. In the swarm behavior algorithm, phenomenological differences play an important role in coordinating swarm behavior. The present invention uses the LED blinking frequency to transmit the difference information between individuals. Each robot can transmit its own state information through five different blinking modes. The specific implementation steps are as follows: The robot monitors the LED brightness status within the State-ROI area of its neighbors to determine whether blinking occurs.

[0037] If the State-ROI brightness status changes between two adjacent frames, it is considered that the status of the neighbor has changed, and the corresponding status information is transmitted to other members in the swarm.

[0038] The robot adjusts its behavior according to the blinking frequency information of its neighbors, thereby achieving collaborative operations of the group.

[0039] S3. Verification and Experiment of the Clustering Algorithm

[0040] Experimental System and Verification Environment. The performance test of the implicit collaboration system relies on an indoor motion capture system to obtain ground truth data. This system combines multiple cameras with markers and uses sensors and software technology to capture the motion postures of objects or humans. Its basic principle is to identify the high-reflectivity markers on the object through the camera, record the motion trajectory, and calculate the pose information of the tracked object based on geometric relationships, and finally convert it into a mathematical model and computer image. This technology has been widely applied in multiple fields such as robotics, sports, and healthcare. There are various types of markers, including objects such as reflective balls and LED lights that are easily captured by the camera.

[0041] In this experiment, the true motion state of the observed drone is obtained through the motion capture system, and it is compared and verified with the relative motion state estimation results obtained by the on-board sensors. The specific steps are as follows: Test Area and Equipment Preparation: Select an open area to avoid obstacles in the field of view, and place at least three markers to facilitate the system to solve the position and pose of the robot in three-dimensional space.

[0042] Motion Capture System Calibration: Ensure that the positions and angles of all cameras are accurate and verify them through markers at known positions.

[0043] Data Collection and Analysis: Place the drone to be tested in the test area, start the multi-camera system for pose calculation, record the true global position and pose of the drone, generate a ground truth data set, and use statistical analysis tools for data analysis.

[0044] During the experiment, the following key points need to be noted: Ensure that the markers are not blocked and the cameras can capture each marker completely. Use high-reflectivity and easily captured markers such as reflective balls or LED lights. Try to avoid interference factors such as wind and light in the test environment. Select an appropriate data collection frequency according to the motion ability of the robot.

[0045] The experiment also tests the real-time performance of the perception system to ensure that the robot can meet the real-time update requirements of the cluster model. The default single-camera configuration is used in the experiment, and the Jetson Nano main control board is used to provide computing power.

[0046] The actual execution time of each algorithm module is tested. The experimental results show that the average execution time of the function modules such as neighbor recognition, identity matching, and cooperative localization of the implicit collaboration system meets the actual application requirements of the cluster algorithm. The update frequency of the implicit collaboration system is 32Hz, and the performance is excellent.

[0047] Verification and analysis of the cooperative positioning accuracy of multiple neighbors. The cooperative positioning accuracy of the robot was tested through experiments under different relative distances and azimuth angles. The experimental results show that in most cases, the robot can keep the ranging error rate within 6% and the error rate of the relative azimuth angle within 8%, proving that the ICS algorithm has high accuracy in cooperative positioning.

[0048] Verification and analysis of the measurement accuracy of the relative velocity direction of multiple neighbors. In the experiment, the accuracy of measuring the relative velocity direction of the implicit cooperation system was tested under different distances and azimuth angles. The results show that in more than 90% of the cases, the implicit cooperation system can control the relative velocity direction error within 20°, and in 42.6% of the cases, the error is controlled within 5°.

[0049] Verification and analysis of the measurement accuracy of phenomenological differential information. The experimental results show that the implicit cooperation system can measure the phenomenological differential information of neighbors relatively accurately with a low error rate, meeting the verification requirements of the clustering algorithm. The main sources of error are factors such as image quality and motion blur.

Claims

1. A cluster robot platform based on implicit collaboration and multi-dimensional perception, characterized in that Including: Robot hardware system, the hardware system includes: Underlying motion system, adopting a two-wheel differential motion scheme, including stepper motor drive wheels on both the left and right sides and two universal wheels in the front and rear. The motion system can accurately control the movement of the robot according to the requirements of the clustering algorithm; Optical cooperation beacon system, including multiple LED beacons. The LED beacons are distributed on the positioning rod and the heading rod, and are used to provide morphological information and positioning information of the robot, and help the robot identify the positions and relative orientations of itself and its neighbors by emitting optical signals; Main control board, the main control board communicates with other robots through a wireless module and controls the robot to execute the clustering algorithm; Camera system, used to collect real-time images of the surrounding environment. The camera has a 160° viewing angle and can detect and identify the states and relative positions of neighboring robots; Robot software system, used to control the clustering behavior of the robot. The software system includes: Multi-neighbor recognition and identity matching mechanism, which detects and pairs neighboring robots through image processing technology and uses the geometric center coordinates of the blue light spot for neighbor recognition; Cooperative positioning mechanism, which calculates the relative distance and azimuth angle of neighbors by calculating the optical characteristics of neighboring robots; Relative speed direction measurement mechanism, which calculates and measures the relative speed direction of neighbors based on the optical characteristics of the heading rod of neighbors; Phenomenological differential information perception mechanism, which transmits information through the blinking frequency of the LED, detects changes in the states of neighbors and conducts clustering coordination.

2. The swarm robot platform according to claim 1, wherein, The motion system uses stepper motors instead of brushed DC motors. The stepper motors provide high-precision speed control, ensuring that the robot can move smoothly at a low speed, and has high response accuracy, meeting the requirements of the clustering algorithm.

3. The swarm robot platform according to claim 1, wherein, The optical cooperation beacon system includes at least two LED groups, which are respectively used to provide positioning information and heading angle information of the robot. The LED groups transmit signals by emitting infrared light and blue light respectively according to different functional uses, and the signals can be evenly transmitted in all 360° directions within a range of 2 meters.

4. The swarm robot platform according to claim 1, wherein, The camera system includes at least one camera and can be configured as a dual camera according to needs to provide a wider field of view and increase the perception range of the clustering robots.

5. The swarm robot platform according to claim 1, wherein, The software system realizes multi-neighbor recognition and identity matching through image processing technology, uses HSV color space conversion and contour extraction algorithms to accurately identify neighboring robots, and realizes identity matching based on geometric analysis.

6. The swarm robot platform according to claim 1, wherein, The cooperative positioning mechanism adopts a simplified image feature point-based positioning method, calculates the relative positions of neighboring robots through the position relationship of LED light spots, and calculates the relative azimuth in combination with geometric relationships.

7. The swarm robot platform according to claim 1, wherein The relative speed direction measurement mechanism calculates the relative speed direction of neighbors by distinguishing the optical characteristics of the forward and rearward heading rods of neighboring robots, and provides high-precision speed direction measurement according to the proportional relationship between the pixel distance and the actual distance in the visual image.

8. The swarm robot platform according to claim 1, wherein, The described phenomenological differential information perception mechanism is based on the LED blinking frequency, and transmits the robot state information through the change of the blinking frequency, helping the swarm robots to coordinate swarm behavior and improving the efficiency and stability of the swarm.

9. The swarm robot platform according to claim 1, wherein, The described system supports the verification and adaptation of multiple swarm algorithms, including swarm algorithms based on inter-neighbor distance, velocity consistency, and phenomenological differences. The platform can effectively integrate and verify different swarm algorithms to achieve intelligent swarm behavior.

Images