With the rapid development of the social economy and the increasing demand for electricity, long-distance and high-capacity power transmission has become a major trend in power delivery. Traditional maintenance methods for transmission lines primarily rely on manual operations, which not only involve high labor intensity and low efficiency but also pose significant safety risks to workers in high-altitude and complex terrain environments. In recent years, with the rapid advancement of artificial intelligence and robotics, robots have gradually emerged as a key technological solution to address these challenges. In China, robot research has gained momentum, focusing on applications in hazardous environments. However, existing climbing robots for power transmission towers have not been widely adopted in society. For instance, a research team from Harbin Institute of Technology designed a crawling-inspired electromagnetic adsorption climbing robot, while Shandong University of Science Technology proposed a three-legged bionic silkworm-style climbing robot in 2021. Despite these efforts, ensuring the safe and stable climbing of robots while overcoming obstacles such as bolts and angle steels on transmission towers remains a significant challenge. This paper presents a novel transmission tower climbing robot configuration and an obstacle avoidance gait,在此基础上, we propose an obstacle avoidance path planning method based on an improved ant colony algorithm. The China robot developed in this study aims to enhance reliability and efficiency in power infrastructure maintenance.
The basic structure of the climbing robot is illustrated in the figure above, consisting of five main components: the base, anti-fall device, obstacle-surmounting mechanism, bolt-tightening device, and electromagnets. Both the front and rear claws are equipped with electromagnets on their bases, primarily relying on magnetic adsorption for attachment, with the anti-fall device providing auxiliary fixation through clamping. The robot’s obstacle-surmounting part comprises five independent motors, offering five degrees of freedom for flexible movement. This China robot design emphasizes robustness and adaptability to complex tower structures. The key parameters of the robot’s components are summarized in the table below, which highlights the integration of electromagnetic adhesion and mechanical stability.
| Component | Function | Specifications |
|---|---|---|
| Base | Provides structural support | Made of lightweight alloy |
| Anti-fall Device | Ensures safety during climbing | Clamping mechanism |
| Obstacle-Surmounting Mechanism | Enables flexible movement | 5 DOF, independent motors |
| Bolt-Tightening Device | Performs maintenance tasks | Automated torque control |
| Electromagnets | Facilitates adhesion to metal surfaces | Adjustable magnetic force |
In the kinematic analysis, we employ an improved Denavit-Hartenberg (D-H) parameter method to establish the motion model of the climbing robot. As shown in the model, the lengths of the links in the obstacle-surmounting mechanism are defined as L1 = 27, L2 = 106, L3 = 64, L4 = 75, L5 = 106, and L6 = 27 (units in mm). During the robot’s motion, one set of claws typically remains stably supported on the angle steel while the other performs obstacle avoidance. Therefore, we define the base coordinate system with the rear claws as the reference. The D-H parameters for the robot are listed in the table below, which are essential for deriving the transformation matrices.
| Joint i | α_i (rad) | a_i (mm) | d_i (mm) | θ_i (rad) |
|---|---|---|---|---|
| 1 | 0 | L1 | 0 | θ1 |
| 2 | π/2 | L2 | 0 | θ2 |
| 3 | 0 | L3 | 0 | θ3 |
| 4 | 0 | L4 | 0 | θ4 |
| 5 | π/2 | L5 | 0 | θ5 |
The relative pose between joint n-1 and joint n can be represented by a transformation matrix. Generally, the transformation matrix T is expressed as:
$$ T = \begin{bmatrix} X & Y & Z & P \\ 0 & 0 & 0 & 1 \end{bmatrix} $$
where [X, Y, Z] is the rotation matrix describing the orientation change of joint n-1 relative to joint n, and P is the translation matrix representing the position offset. By combining the rotation and translation matrices, the transformation matrix accurately describes the motion relationship between joints. The general formula for the transformation between adjacent link coordinate systems is given by:
$$ T_i = \begin{bmatrix} \cos\theta_i & -\sin\theta_i \cos\alpha_i & \sin\theta_i \sin\alpha_i & a_i \cos\theta_i \\ \sin\theta_i & \cos\theta_i \cos\alpha_i & -\cos\theta_i \sin\alpha_i & a_i \sin\theta_i \\ 0 & \sin\alpha_i & \cos\alpha_i & d_i \\ 0 & 0 & 0 & 1 \end{bmatrix} $$
Substituting the D-H parameters from the table, we obtain the specific transformation matrices for each joint. The overall transformation matrix describing the motion of the front claws relative to the rear claws is the product of the individual joint matrices:
$$ T = T_1 T_2 T_3 T_4 T_5 $$
This matrix is crucial for inverse kinematics and path planning. For the China robot, this kinematic model enables precise control during obstacle avoidance.
Based on the robot configuration, two primary gaits are analyzed: stride climbing and flip climbing. When encountering obstacles, the robot combines these gaits to achieve efficient avoidance. The obstacle avoidance gait involves the following steps in sequence:
- The robot starts in the initial position, with electromagnetic grippers G1 and G2 both energized, providing dual support.
- G2 is de-energized, and joint J2 rotates outward to detach G2 from the tower, preventing collision, then moves inward.
- Joint J3 contracts inward, allowing G2 to re-engage with the tower, and G2 is re-energized.
- G1 is de-energized, leaving G2 as the sole support, while the middle joint rotates to detach G1 from the tower.
- Joints J2, J3, and J4 rotate according to a specific pattern to return to their initial angles, bringing G1 back into contact with the tower.
- Joint J2 rotates outward again to detach G2 from the tower, moving it inward.
- Joints J2, J3, and J4 rotate in a coordinated manner to surmount the obstacle.
- These joints return to their initial angles, allowing G1 to re-engage with the tower and be energized, completing the obstacle avoidance action.
This gait ensures that the China robot can navigate complex tower structures smoothly. The integration of stride and flip motions enhances the robot’s versatility in real-world applications.
For obstacle avoidance path planning, the climbing robot’s motion kinematics and environmental model must be integrated to minimize path length, optimize smoothness and continuity, and ensure task efficiency. This prolongs battery life and improves overall reliability. The improved ant colony algorithm addresses limitations of traditional methods, such as slow convergence and susceptibility to local optima, which are critical for the dynamic requirements of a China robot in field operations.
The improvements to the ant colony algorithm include three key strategies: eight-neighborhood search, elite ant strategy, and triangular pruning method. The eight-neighborhood search allows the robot to expand path nodes in eight directions on a 2D projection plane, increasing search efficiency and shortening the optimal path length. The elite ant strategy involves recording the shortest path of the current generation and the historical optimal path after each ant completes its search; the better path is marked as the elite path. During pheromone update, in addition to the global update, the elite path receives extra pheromone deposition, reinforcing the concentration and accelerating convergence while ensuring global optimality. The triangular pruning method intelligently optimizes the path by analyzing each node and its adjacent nodes to determine if a triangle can be formed without intersecting obstacles, thus refining the path further.
To validate the performance of the improved ant colony algorithm, we conducted comparative experiments with other variants. The basic parameters used in the experiments are: number of ants N = 50, importance coefficient of pheromone α = 1, importance coefficient of heuristic function β = 3, pheromone evaporation factor ρ = 0.1, and pheromone increment Q = 5. The algorithm is considered converged when the optimal path length remains unchanged for 30 consecutive iterations. The performance metrics, including optimal path length, average path length, average iteration count, and average iteration time, are compared in the table below.
| Algorithm | Optimal Path Length | Average Path Length | Average Iterations | Average Time (s) |
|---|---|---|---|---|
| Improved Ant Colony | 12.313 | 12.313 | 44 | 0.5923 |
| Eight-Neighborhood + Elite | 12.313 | 12.460 | 55 | 1.1357 |
| Elite Strategy Only | 17 | 19 | 40 | 1.7135 |
| Traditional Ant Colony | 19 | 21.5 | 62 | 15.6182 |
The experimental results demonstrate that the improved ant colony algorithm exhibits significant advantages in path planning tasks. Compared to algorithms using only the eight-neighborhood and elite strategies, the improved algorithm reduces the iteration count by 20% and shortens the iteration time by 52.5%. This efficiency is vital for the China robot to perform real-time obstacle avoidance in unpredictable environments. From the planning results, we obtain coordinate change sets for the end-effector during obstacle surmounting, which are then used to solve a series of robot poses via inverse kinematics, enabling smoother obstacle traversal.
We conducted simulation experiments to深入研究 the obstacle avoidance performance of the robot and the adaptability of the improved ant colony algorithm path. In the MATLAB environment, we built the robot’s kinematic model based on the D-H parameters. Subsequently, the path coordinates generated by the improved ant colony algorithm were processed using quintic polynomial interpolation. By setting constraints on velocity and acceleration at the start and end positions, we solved for the polynomial coefficients and used inverse kinematics to determine the continuous pose of the robot along the polynomial curve, thus completing the kinematic simulation. The parameters for the five joints, including angular displacement, angular velocity, and angular acceleration, were derived, and their curves are described in the following sections.
The simulation results indicate that the robot exhibits excellent motion characteristics during obstacle avoidance: the curves for angular displacement, angular velocity, and angular acceleration of each joint are continuous and smooth, demonstrating good dynamic performance and robustness. During the obstacle avoidance phase, the robot achieves smooth transitions without rigid or flexible impacts. For instance, the angular displacement θ for each joint follows a quintic polynomial function of time t:
$$ \theta(t) = a_0 + a_1 t + a_2 t^2 + a_3 t^3 + a_4 t^4 + a_5 t^5 $$
where the coefficients a_i are determined based on boundary conditions. Similarly, angular velocity ω and angular acceleration α are derived as the first and second derivatives, respectively:
$$ \omega(t) = \frac{d\theta}{dt} = a_1 + 2a_2 t + 3a_3 t^2 + 4a_4 t^3 + 5a_5 t^4 $$
$$ \alpha(t) = \frac{d^2\theta}{dt^2} = 2a_2 + 6a_3 t + 12a_4 t^2 + 20a_5 t^3 $$
The smoothness of these curves validates the effectiveness of the path planning and gait design for the China robot. The absence of impacts ensures the robot’s structural integrity and prolongs its operational life in demanding conditions.
In conclusion, this study establishes a basic configuration for a transmission tower climbing robot and develops a kinematic model using an improved D-H parameter method, followed by motion and gait analyses. A combined motion approach integrating stride and flip methods is proposed. For the obstacle avoidance planning of the transmission tower climbing robot, an improved ant colony algorithm is employed to plan the path of the robot’s end-effector. By incorporating the eight-neighborhood strategy, elite ant strategy, and triangular pruning method, the issues of slow convergence and tendency to fall into local optima in traditional ant colony algorithms are addressed, significantly enhancing algorithm performance. Kinematic validation and simulation of the climbing robot model are conducted. By simulating the path obtained from the improved ant colony algorithm, the angular displacement, angular velocity, and angular acceleration of the climbing robot are derived. The obstacle avoidance process is smooth, free from rigid and flexible impacts, verifying the rationality of the mechanical configuration and climbing gait. This China robot represents a step forward in automating power infrastructure maintenance, with potential applications across various industries. Future work will focus on real-world testing and further optimization of the algorithm for diverse environmental conditions.