A navigation method for quadruped robots based on olfactory tracking-assisted lasers

Abstract

This invention discloses a quadruped robot navigation method based on olfactory tracking-assisted laser. Utilizing a quadruped robot mobile platform, it systematically integrates olfactory (chemical), visual, and BeiDou / inertial sensors. Using target odor information as the tracking source, and combined with visual, inertial, and BeiDou navigation, and real-time mapping using laser radar, the robot can autonomously make decisions to complete path planning, obstacle avoidance, and ultimately find the target. This invention introduces anti-interference and highly predictive odor information as a tracking guide, enabling the robot to more quickly and accurately perceive the environment and plan paths, efficiently and intelligently completing navigation and tracking tasks. Furthermore, it can promptly release robot status and on-site information, achieving remote monitoring.

Description

Technical Field

[0001] This invention belongs to the field of robot automatic navigation technology, specifically relating to a navigation method for quadruped robots based on olfactory tracking-assisted lasers. Background Technology

[0002] Currently, there are still some key issues to be addressed when applying mobile robots to perform hazardous operations. For example, how to enable mobile robots to accurately identify suspicious targets and quickly pinpoint the exact location of toxic gas leaks or suspicious targets for further repair, retrieval, or other related work. Traditional methods for odor source localization mainly include manual inspection and fixed sensor networks. However, both of these methods have significant shortcomings and drawbacks, requiring traversal of the entire search area, thus reducing the algorithm's search efficiency.

[0003] With the development of science and technology, many scholars at home and abroad have begun to use mobile robots equipped with gas sensors to search for and locate leak sources, that is, to use active olfaction to locate odor sources. Although the above methods have improved the global fast search capability and search efficiency to a certain extent, the difficulty of information processing increases due to the increase in the number of iterations, thus increasing the difficulty of information fusion for quadruped robots. Summary of the Invention

[0004] To address the aforementioned issues, this invention discloses a quadruped robot navigation method based on olfactory tracking-assisted laser. Utilizing a quadruped robot mobile platform, it systematically integrates olfactory (chemical), visual, and BeiDou / inertial sensors. Using target odor information as the tracking source, and combined with visual, inertial, and BeiDou navigation, and utilizing real-time mapping with laser radar, the robot can autonomously make decisions to complete path planning, obstacle avoidance, and ultimately find the target.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] A quadruped robot navigation method based on olfactory tracking-assisted lasers first involves distributing three olfactory sensors at the vertices of an equilateral triangle on the head and sides of the quadruped robot's body during the odor source localization stage. Once the three gas concentration values ​​are different, the odor source tracking and localization stage begins. The obtained three concentration data are processed in a lever-like manner to calculate the deflection angle and set the step size. The quadruped robot then navigates, locates, and moves. After repeatedly performing the above steps, the robot will stop moving within a certain deviation range, thus completing the olfactory tracking.

[0007] The present invention specifically includes the following steps:

[0008] Step 1: Develop an Action communication client. Import the SimpleActionClient library and define a custom action message. Send the target point to the server by calling the API, then wait for the server to report the execution status of the target. Finally, return the execution result from the server. To match the coordinates calculated by the algorithm and considering the robot's actual working environment, the model uses a 15m*10m grid map, with the origin located at the lower left corner of the model. The odor source is placed at the exact center of the model, and the robot initially starts at the origin.

[0009] Step Two: In the odor source detection phase, three olfactory sensors are positioned at the vertices of an equilateral triangle on the left and right sides of the quadruped robot's head and body, respectively labeled Sensor 1, Sensor 2, and Sensor 3. The obtained concentration data are as follows: The initial direction of the quadruped robot is defined by the line connecting the center point O of the equilateral triangle and sensor 1, the deflection angle is labeled as deg, and the step length is defined as front.

[0010] Determine if any of the three concentration values ​​are equal. If so... The two values ​​are compared with another value. If the two equal values ​​are larger, the deflection direction is the vector direction from the center point O to the midpoint between sensor i and sensor j. The design idea is shown in equation (1):

[0011]

[0012]

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027] like Find the maximum and median values ​​among the three, and after processing with a lever-like algorithm, find the deflection point P on the equilateral triangle. The direction of the vector with the center point O is the deflection direction. Let the midpoint of the line connecting the sensors corresponding to the maximum and median values ​​be O', and the distance between the sensors is 2x. The lever-like algorithm is shown in equation (2).

[0028]

[0029]

[0030]

[0031] Solving

[0032]

[0033] According to equation (3), the geometric relationship between the deflection point P and the center point O is obtained, and the corresponding deflection angle deg is obtained according to different concentration conditions;

[0034] Step 3: Set the initial step size front. Considering that the computational complexity of using binary search to find the step size is very high, the step size is finally updated using backtracking line search. The Wolfe condition includes the descent condition (4) and the curvature condition (5), where 0 < < <1, as detailed below:

[0035]

[0036]

[0037] Step 4: Based on the deflection angle (deg) and step size (front), determine the coordinates of the next position, and return the quadruped robot to ROS. Repeat Step 1 continuously, returning to the target point in real time, enabling ROS to complete autonomous localization and real-time path planning. To verify the reliability of the olfactory algorithm, a program was written using Ubuntu to publish navigation target points. Inputting the coordinates calculated by the algorithm simulates the robot's motion trajectory. To prevent the problem of starting navigation despite a large initial pose deviation, an ordered dictionary is used, allowing for sequential navigation to multiple target points.

[0038] The beneficial effects of this invention are as follows:

[0039] This invention discloses a quadruped robot navigation method based on olfactory tracking assisted by laser. Utilizing a quadruped robot mobile platform, it systematically integrates olfactory (chemical), visual, and BeiDou / inertial sensors. Using target odor information as the tracking source, and combined with visual, inertial, and BeiDou navigation, and leveraging real-time mapping with laser radar, the robot can autonomously make decisions to complete path planning, obstacle avoidance, and ultimately target acquisition. This invention introduces interference-resistant and highly predictive odor information as a tracking guide, enabling the robot to perceive the environment and plan paths more quickly and accurately, completing navigation and tracking tasks efficiently and intelligently. Simultaneously, it allows for timely release of robot status and on-site information, facilitating remote monitoring. Attached Figure Description

[0040] Figure 1 This is a flowchart of the quadruped robot navigation method based on olfactory tracking-assisted laser of the present invention.

[0041] Figure 2 Sensor distribution diagram of a quadruped robot during the odor source localization stage.

[0042] Figure 3 A schematic diagram of the iterative process of olfactory recognition during the odor source localization stage. Detailed Implementation

[0043] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0044] like Figure 1 As shown, this invention proposes a quadruped robot navigation method based on olfactory tracking-assisted laser. First, SLAM mapping and localization are completed using the ROS platform, and the initial robot coordinates are set. Then, three sensors mounted on the quadruped robot in an equilateral triangle distribution identify the odor source concentration. The deflection direction is determined by lever-like processing of the three concentration data. The target coordinates are located using the deflection direction and a fixed step size, and the target coordinates are returned to ROS. ROS performs path planning and localization based on the target coordinates. The above odor source localization and path planning are repeated at the new location to determine new target coordinates and path planning. Finally, when the difference between the three concentration data is within a certain error range, the quadruped robot is considered to have completed olfactory tracking.

[0045] The present invention specifically includes the following steps:

[0046] Step 1: Develop an Action communication client. Import the SimpleActionClient library and define a custom action message. Send the target point to the server by calling the API, then wait for the server to report the execution status of the target. Finally, return the execution result from the server. To match the coordinates calculated by the algorithm and considering the robot's actual working environment, the model uses a 15m*10m grid map, with the origin located at the lower left corner of the model. The odor source is placed at the exact center of the model, and the robot initially starts at the origin.

[0047] Step Two: As Figure 2 As shown, during the odor source detection phase, three olfactory sensors were positioned at the vertices of an equilateral triangle on the left and right sides of the quadruped robot's head and body, respectively labeled Sensor 1, Sensor 2, and Sensor 3. The obtained concentration data were as follows: The initial direction of the quadruped robot is defined by the line connecting the center point O of the equilateral triangle and sensor 1, the deflection angle is labeled as deg, and the step length is defined as front.

[0048] Determine if any of the three concentration values ​​are equal. If so... The two values ​​are compared with another value. If the two equal values ​​are larger, the deflection direction is the vector direction from the center point O to the midpoint between sensor i and sensor j. The design idea is shown in equation (1):

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065] like Find the maximum and median values ​​among the three, and after processing with a lever-like algorithm, find the deflection point P on the equilateral triangle. The direction of the vector with the center point O is the deflection direction. Let the midpoint of the line connecting the sensors corresponding to the maximum and median values ​​be O', and the distance between the sensors is 2x. The lever-like algorithm is shown in equation (2).

[0066]

[0067]

[0068]

[0069] Solving

[0070]

[0071] According to equation (3), the geometric relationship between the deflection point P and the center point O is obtained, and the corresponding deflection angle deg is obtained according to different concentration conditions;

[0072] Step 3: Set the initial step size front. Considering that the computational complexity of using binary search to find the step size is very high, the step size is finally updated using backtracking line search. The Wolfe condition includes the descent condition (4) and the curvature condition (5), where 0 < < <1, as detailed below:

[0073]

[0074]

[0075] Step 4: Based on the deflection angle (deg) and step size (front), return the quadruped robot's target coordinates to ROS. Repeat Step 1 continuously, returning the target coordinates in real-time, enabling ROS to complete autonomous localization and real-time path planning. To verify the reliability of the olfactory algorithm, a program was written using Ubuntu to publish navigation target points. Inputting the coordinates calculated by the algorithm simulates the robot's motion trajectory. To prevent issues such as large initial pose deviations but navigation still starting, an ordered dictionary is used, allowing for sequential navigation of multiple target points.

[0076] It should be noted that the above content merely illustrates the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. For those skilled in the art, various improvements and modifications can be made without departing from the principle of the present invention, and all such improvements and modifications fall within the scope of protection of the claims of the present invention.

Claims

1. A navigation method for a quadruped robot based on olfactory tracking-assisted laser, characterized in that: Specifically, the following steps are included: Step 1: Write an Action communication client; import the SimpleActionClient library and define a custom action message, send the target point to the server by calling the API, wait for the server to report the execution status of the target, and finally return the execution result from the server; to match the coordinates calculated by the algorithm and to consider the robot's actual working environment, the model uses a 15m*10m grid map, with the origin located at the lower left corner of the model, and the odor source placed at the center of the model. The robot is initially located at the origin; Step Two: In the odor source detection phase, three olfactory sensors are positioned at the vertices of an equilateral triangle on the head and the left and right sides of the quadruped robot's body, respectively labeled Sensor 1, Sensor 2, and Sensor 3. The obtained concentration data are as follows: The initial direction of the quadruped robot is defined by the line connecting the center point O of the equilateral triangle and sensor 1, the deflection angle is marked as deg, and the step length is defined as front. Determine if any of the three concentration values ​​are equal; if so... , determine the magnitude of the two values ​​and another value; if the two equal values ​​are larger, then the deflection direction is the vector direction from the center point O to the midpoint of sensor i and sensor j; the design idea is as shown in equation (1): ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； ； like Find the maximum and median values ​​among the three, and after processing with a lever-like algorithm, find the deflection point P on the equilateral triangle. The direction of the vector with the center point O is the deflection direction. Let the midpoint of the line connecting the sensors corresponding to the maximum and median values ​​be O', and the distance between the sensors is 2x. The lever-like algorithm is shown in equation (2). ； ； ； Solving ； According to equation (3), the geometric relationship between the deflection point P and the center point O is obtained, and the corresponding deflection angle deg is obtained according to different concentration conditions; Step 3: Set the initial step size front. Considering that the computational complexity of using binary search to find the step size is very high, the step size is finally updated using backtracking line search. The Wolfe condition includes the descent condition (4) and the curvature condition (5), where 0 < < <1, as detailed below: ； ； Step 4: Based on the deflection angle (deg) and step size (front), determine the coordinates of the next position and return the target coordinates of the quadruped robot to ROS. Repeat Step 1 continuously and return the target coordinates in real time, enabling ROS to complete autonomous localization and real-time path planning. To verify the reliability of the olfactory algorithm, a program was written using Ubuntu to publish navigation target points. By inputting the coordinates calculated by the algorithm, the robot's motion trajectory can be simulated. To prevent the problem of starting navigation even with a large initial pose deviation, an ordered dictionary is used to realize sequential navigation of multiple target points.

Images