Abstract : This paper first explores the application strategies and required functions of robots for rescue or assisted rescue. It studies the unstructured terrain environment that needs to be overcome, extracts and simplifies its features, and proposes performance requirements and indicators for the mechanical system of a coal mine rescue robot. Fractal interpolation and cubic spline interpolation methods are used to simulate three-dimensional real terrain, which is beneficial for the dynamic simulation and optimization of its mechanical system. A coal mine rescue robot is designed using a four-tracked double-swing-arm walking mechanism, and its practical application is analyzed.
Keywords : mine rescue robot; feature extraction; unstructured terrain; four-tracked double-swing arm type
1 Introduction doubleswingarmtype
my country has abundant coal resources and is the world's largest producer and consumer of coal. Coal accounts for approximately 70% of my country's primary energy production and consumption. Currently, my country's rapid economic growth places higher demands on the development of the coal industry.
In recent years, coal mine accidents have occurred frequently in China. Among these accidents, fire and explosion-related accidents account for a considerable proportion, and the main substances causing fires and explosions in mines are gas and dust. The complex geological conditions of coal seams and the many factors affecting safe production in coal mines are objective factors that cause accidents. Most of the coal seams mined in China have high gas content, low permeability, and complex geological structures, making it difficult to drain gas before mining. Among large and medium-sized coal mines, high-gas mines account for 20.34 %, and gas outburst mines account for 19.77 %. Among small coal mines, high-gas mines account for about 15% [4]. Once a gas explosion occurs in a mine, the personnel and property underground are in extreme danger and rescue must be organized as soon as possible. During the coal mine rescue process, accidents causing injury or death to rescue personnel due to improper command, violations of regulations, equipment limitations, and secondary explosions also occur frequently.
In summary, due to the geological characteristics of my country's coal mines and some management factors, mine accidents occur frequently. After a mine accident, rescue teams must risk entering the disaster area to explore and rescue those in distress. However, conducting underground rescue and disaster management after a disaster is extremely dangerous. There is an urgent need to develop coal mine rescue robots to replace or partially replace rescue team members in entering mine disaster sites for environmental exploration and search and rescue missions. Therefore, research on coal mine rescue robots has significant practical and social implications for coal mine safety production and disaster relief, and is of great importance for establishing an industrial disaster relief system in special hazardous environments.
2. Modeling of unstructured terrain in underground wells
2.1 Rescue Strategies for Coal Mine Rescue Robots
Gas explosions have a wide impact and cause extensive damage, resulting in numerous casualties. Furthermore, disaster relief is extremely difficult and dangerous. In the initial stages of rescue and relief efforts, the specific extent of the affected area, the location of the explosion source, the concentration of gas and CO, and the extent of damage to the tunnels are almost entirely unknown. Rescue personnel face the dangers of secondary explosions, CO poisoning, the possibility of roof collapse, and the threat of extreme heat. Robots can enter the disaster area before rescue personnel, detect and transmit environmental information. The rescue strategy for coal mine disaster relief robots is as follows:
(1) Cut off the power supply in the disaster area, notify the mine rescue team, quickly establish a disaster relief command center, and set up several rescue teams to perform their respective duties in strict accordance with the requirements of the disaster prevention and treatment plan.
(2) Determine the scope of the disaster area and establish an underground disaster relief command center in a safe location as close as possible to the disaster area, ensuring smooth communication between the above-ground and underground command centers. The location of the underground disaster relief command center should meet the requirements of normal ventilation, gas and CO concentrations within safe ranges, and the ability to block the shock wave of a secondary gas explosion.
(3) Rescue personnel, carrying essential rescue equipment and disaster relief robots, advance towards the disaster area from the underground disaster relief command center. The disaster relief robots can serve as the vanguard of the rescue personnel, entering the disaster area tunnels to detect environmental information such as gas and CO concentrations and temperature within the tunnels.
(4) Multiple coal mine disaster relief robots were used to simultaneously detect suspicious roadways and working faces in the core area of the disaster zone.
2.2 Extraction and Simplification of Unstructured Topographic Environmental Features in Coal Mines
The underground terrain environment of coal mines is a unique unstructured environment. The layout of underground roadways, the construction of tracks and walkways, and the arrangement of equipment are designed, constructed, and installed according to certain standards based on the geological characteristics of the coal mine and the needs of coal production. Therefore, some specific terrain features with fixed shapes and structures exist underground in coal mines, such as track beds and steel rails. The terrain environment and travel length of the mobile robot are the initial conditions for the design of the robot's mechanical system, determining the mechanical performance indicators of the coal mine rescue robot, and further determining the selection of the robot's walking mechanism form and the determination of its dimensions. A practical model conforming to the characteristics of underground coal mines is established for robot simulation; and based on the survey of coal mine disaster areas, a certain degree of unstructured terrain environment is established underground, and computer simulation is conducted.
The tunnel cross-section includes the cross-sectional shape, width, and height of the machine tunnel. This determines the choice of the robot's width. The net width of an arched double-track tunnel is calculated using the following formula.
B = a + 2A1 + C + t ≥ 2.4m (2-1)
In the formula, B is the net width of the roadway, referring to the horizontal distance between the inner sides of the straight end; a is the width of the non-pedestrian side as specified in the "Coal Mine Safety Regulations," a≥0.3m; when a conveyor is installed in the roadway, the distance between the conveyor and the most protruding part of the support or arch wall, a≥0.5m. A1 is the maximum width of the conveying equipment. C is the width of the pedestrian walkway as specified in the "Coal Mine Safety Regulations," within a height of 1.6m from the roadway ballast surface, C≥0.8m, and at pedestrian parking locations, C≥1.0m. Pipes, lines, and cables are not allowed to be installed between 1.6m and 1.8m.
Figure 1 Track cross section
2.3 Computer Modeling of 3D Feature Terrain and Real Terrain
The dimensional parameters of the underground coal mine terrain environment were investigated, and the terrain features were extracted and simplified. Three-dimensional mapping software was used to model the three-dimensional terrain features. The height of independent steps in underground coal mines is mostly below 250mm. The height h of continuous steps is generally 150~180mm, the step span b is generally 220~350mm, and the step width B is generally...
The slope of the continuous steps, ranging from 600 to 1200 mm, often matches the slope of the roadway, typically between 8 and 15 degrees. Based on the simplified features and dimensions of the underground coal mine terrain, a three-dimensional feature terrain model was created. Figure 2 shows one such three-dimensional feature terrain. This model comprises four main parts: a raised area, a stepped area, a channel area, and a slope area.
Figure 2 Three-dimensional feature terrain
2.3 Computer Simulation and Modeling of 3D Realistic Terrain
Extracting and simplifying features of unstructured terrain in coal mines is beneficial for simulating and modeling obstacle-crossing motion of robot virtual prototypes on terrain with characteristic obstacles.
2.3.1 Spline interpolation simulation for complex terrain
Digital elevation model (DEM) is a discrete digital representation of the topography of the ground environment, represented by the Gaussian coordinates (x, y) of grid points and their corresponding elevations Z [5][7]. The three-dimensional display of the terrain first requires the acquisition of DEM data. For small areas of ground, a three-dimensional laser scanner can be used to acquire its DEM data. Due to the limitations of objective conditions, it is often impossible to obtain enough sampling points to meet the needs of display and simulation. Therefore, interpolation is required to generate more points. Commonly used interpolation methods include inverse distance weighted interpolation, bilinear interpolation, trend surface interpolation, and spline interpolation. MATLAB provides two-dimensional interpolation commands, and its function format is: Z=interp2(x , y , z , X , Y , 'method'). Spline interpolation has higher accuracy than linear interpolation and cubic interpolation and is the most widely used in practical applications. Therefore, cubic spline interpolation was chosen for two-dimensional interpolation operations.
A 3500×3600mm undulating terrain area was selected, and a set of orthogonal grids with individual cells of 250×300mm was used. The height at each grid intersection point was measured to obtain a set of regular two-dimensional matrix elevation values for the terrain. The data was then subjected to cubic spline interpolation, with the X and Y parameters of the new network formed by the interpolation points spaced at intervals of 30mm and 60mm, respectively. The resulting three-dimensional surface plot after interpolation is shown in Figure 3.
Figure 3. Surface generated after spline interpolation of the original data.
Research on the Mechanical System and Walking Mechanism Configuration of Mine Rescue Robot
3.1 Study on the configuration of the walking mechanism
Tracked walking mechanisms are an extension of wheeled walking mechanisms, with the tracks themselves serving to continuously pave the way for the wheels. A common type of tracked walking mechanism uses square tracks, as shown in Figure 4. This walking mechanism consists of a suspension system, drive wheels, guide wheels, load-bearing wheels, track support rollers, and tracks. The drive wheels and guide wheels also function as support wheels (load-bearing wheels), providing a large ground contact area and good stability.
Figure 4. Square tracked walking mechanism
Obstacle-crossing mechanism of a four-tracked, double-swing-arm tracked robot
Figure 5. Trajectory of the robot's center of mass.
Establish a coordinate system xo1y as shown in Figure 5, with the rear track wheel axle of the robot as the origin. Let the distance between the front and rear tracks O1O2 of the main track part of the robot be l0 , the mass of the main body be m1 , and the coordinates of the centroid G1 be ( l1 , h1 ). When climbing typical obstacles, the two swing arms need to swing synchronously. Therefore, let the mass of the two swing arms be m2 , the centroid be G2 , located on the swing arm centerline O2O3 , and the length from the front track wheel axle O2 be l2 . Let the distance between the two track wheel axles of the swing arm be l3 , and the swing angle of the swing arm be θ , where θ∈[0,2π]. Let the radius of the main track wheel be R, and the radius of the front track wheel of the swing arm be r, both including the track thickness. Let the width of the robot be b. Then the coordinates of the robot's centroid ( xG , yG ) are: The centroid of the swing arm track robot satisfies the following relationship.
The center of mass of the swing-arm tracked robot satisfies the following relationship:
| The trajectory of the robot's center of mass as the swing angle θ of the swing arm changes is based on... | With center O, and |
A circle with radius .
Figure 6 shows the process of a four-tracked double-arm robot climbing a step in the forward direction. With the help of the initial swing angle of the swing arm, the robot, driven by the track mechanism, makes the front end of its main track rest on the outer corner line of the step. The robot continues to move, driving the swing arm to swing clockwise. When the robot's center of gravity line crosses the outer corner line of the step, the robot flips towards the upper plane of the step with the outer corner line as the branch line, and the robot successfully climbs the step.
Figure 6. The process of the robot climbing the stairs forward.
Once the robot's main track rests against the outer corner of the step, there are three ways to climb it, depending on the step's height:
Firstly, the main track and the swing arm track are aligned, and the robot climbs using a method similar to that of a fixed track, relying on its center of gravity to cross the outer corner line and flip to the upper platform of the step. This process is shown in Figure 4-18A. At this time, the total length of the robot's tracks is...
Secondly, when the height of the step is less than or equal to the length of the swing arm (H ≤ l³ + rR), rotating the swing arm ensures that the robot's main body is stably placed on the upper platform of the step without the swing arm leaving the ground. Once the robot's main body is placed on the upper platform, the robot's motion ensures that its center of mass moves past the outer corner of the step, thus successfully climbing the step, as shown in the figure.
As shown. This climbing method allows the robot to cross obstacles smoothly and avoids the impact of falls. Third, when the height of the step is greater than the length of the swing arm (3H > l + r - R), the robot's center of mass can be moved closer to and across the outer corner of the step by rotating the swing arm, thus completing the step climbing action. The height H that the robot crosses the steps has the following relationship with θ and β: |
The maximum value of H, Hmax, is the maximum step height the robot can climb in reverse. The partial derivative of H with respect to θ is 0.
When climbing stairs, H has a maximum value when the swing angle θ of the swing arm and the tilt angle β of the robot body follow the above relationship. When the mass of the robot swing arm is small and its influence on the position of the robot's center of mass is negligible, the swing arm is in a vertical position, that is, H has a maximum value.
4. Conclusion
This paper addresses the complex and dangerous environment of underground coal mines, particularly after gas and coal dust explosions, where the terrain and explosive gases create highly complex and hazardous conditions. Lack of information hinders timely decision-making and assessment by rescue personnel, leading to the deaths of many trapped individuals due to delayed rescue efforts. Therefore, researching search robots to replace or partially replace rescue personnel for environmental exploration at disaster sites is of significant practical importance for successful rescue operations and minimizing casualties. This paper, using disaster relief as its research background, develops a multi-segment tracked search robot mechanical system and conducts research on key technologies such as search robot dynamics analysis and obstacle-crossing strategies.
References
1. Lin Baiquan, Analysis of the Current Status of Coal Mine Safety in China [N] , China Mining News, 2006.3.3 .
2. Song Yuanming, Ren Jinbiao. Investigation and Research on Underground Disasters in Coal Mines [J]. Mine Safety, 2005. 3.
3. Xie Ying. Feature Extraction and Classification of Unstructured Road Navigation Images [D]. Nanjing: Nanjing University of Science and Technology, 2008. 6
4. Huang Yong'an, Li Wencheng, Gao Xiaoke. Matlab 7.0 /Simulink 6.0 Application Examples Simulation and Efficient Algorithm Development [M]. North
Beijing: Tsinghua University Press, 2008.
5. Ke Zhengyi, He Jianbang, Chi Tianhe Digital Ground Model [M] Beijing: China Science and Technology Press, 1993.
6. Li Yunwang, Ge Shirong, Zhu Hua, Liu Jian. Research on obstacle crossing mechanism and obstacle crossing ability of four-tracked double-swing arm robot [J]. Robot, 2010, 32(2): 157-165.
7. Xie Shunping , Du Jinkang , Luo Weijia , Deng Min. Feature Extraction of Complex Topographic Watersheds Based on DEM [J]. Geographical Research, 2006. 1
8. Sun Bin. Motion Analysis and Control of Portable Ground Mobile Robots [D]. Changsha: National University of Defense Technology, 2006.
9. Xu Shifan. Prospect of Industrial Robots in Mining [J]. Coal Mine Automation, 1993(1): 28-31.
