Abstract : This paper discusses the characteristics of dual-arm robots, elaborates on the current status and development of dual-arm robot research at home and abroad, and puts forward some views on future research of dual-arm robots in light of China's actual situation. Keywords: Dual-arm robot research + reflection 1 Introduction Current industrial robots are designed for the independent operation of single-arm robots. Such robots are only suitable for specific products and working environments and rely on specialized equipment and fixtures. Generally, single-arm robots are only suitable for operating rigid workpieces and are constrained by the environment. With the development of modern industry and the progress of science and technology, single-arm operation is insufficient for many tasks. Therefore, in order to adapt to the increasing complexity of tasks, the increasing intelligence, and the requirements of system compliance, the system has been extended to bimanual coordinated control, that is, two single-arm robots coordinate and cooperate with each other to complete a certain task. However, since the bimanual coordinated control system consists of two robots, they cannot be a simple combination of two single-arm robots. In addition to the control of their common goals, their coordination and control and adaptability to the environment become the key to the combination. Thus, the further application of bimanual coordinated control robot systems is limited. Dual-arm robots can perform functions that are easily achievable by humans. Dual-arm robots are more practical than bi-armed robots. 2. Action and characteristics of dual-arm robots. A dual-arm robot can be compared to two single-arm robots working together. When the influence of other robots is considered an unknown source of interference, one robot is independent of the other. However, as a complete robotic system, there is a dependency between the two arms. They share sensor data, and the two arms are physically coupled through a common connection. Most importantly, the communication between the controllers of the two arms allows one arm to make corresponding actions, trajectory planning, and decisions in response to the other arm's reaction—that is, there is a coordination relationship between the two arms. This can be seen to some extent as similar to the coordinated movements of our human arms. Two single arms in a body correspond to two high-level controllers. Taking the coordination of all movements as a benchmark, the movement process of our own arms includes data sources including complex mechanical systems, bodily feedback, visual feedback, skin contact, slip detection, and brainpower, and uses pre-acquired data to confirm the storage and processing capabilities of this data. This is precisely the key difference between dual-arm robots and a combination of two single-arm robots. The main characteristics of dual-arm robots are as follows: First, they operate without relative motion between the end effector and the arms, such as when lifting heavy objects like steel bars, which is much simpler to control than the corresponding movements of two single-arm robots. Second, even with relative motion between the end effector and the arms, good coordination between the two arms allows for the control and manipulation of flexible objects such as thin plates, something difficult for two single-arm robots to achieve. Third, dual-arm robots can avoid collisions that occur when two single-arm robots work together. Fourth, the two arms of a dual-arm robot can work independently to manipulate and control multiple targets, such as fitting a nut onto a screw. Although dual-arm robots were developed based on single-arm robots, due to the special nature of their function, research results on single-arm robots cannot be simply transferred to dual-arm robots. Therefore, research on dual-arm robots has become a very important area of ​​robotics research. 3. Current Status and Development of Research Abroad Research on dual-arm robots abroad began in the early 1990s, primarily focusing on trajectory planning (including collision avoidance), coordinated control algorithms, and control of manipulating forces or torques. Research on trajectory planning is mainly based on the scenario of multiple robots working in the same environment without collisions. This research is typically divided into path planning and trajectory planning. Some scholars have conducted in-depth research on the dual-arm control problem of robots moving along specific paths. They have established optimal trajectory planning algorithms that consider the robot's dynamic characteristics using the robot's dynamic equations. Furthermore, they have used computer simulation to find the optimal path for each step, thus solving the collision avoidance problem. In summary, these studies have largely solved the two-dimensional trajectory planning problem, but research on three-dimensional space and redundant dual-arm trajectory planning methods and strategies is scarce. Coordinated control of dual-arm robots is a hot topic in dual-arm robot research, and most studies are based on comparing the coordinated control of two single-arm robots working together. Coordinated control of dual arms includes the control of the relative motion between the hand and the work execution position, and the control to ensure the continuity of the target trajectory. Some Japanese scholars, by establishing an evaluation of the operability of the dual-arm system, proposed operational indices for relative operability and relative operational force. Based on these, they designed new control algorithms for a coordinated control system that allows for the separate specification of the working position and the actual target trajectory while meeting the operational indices. Other Japanese scholars, by establishing a hybrid control theory of position and force, used hybrid position and force control experiments to find the optimal joint force or torque to ensure coordinated control of the two arms. Sarkar et al. in the United States proposed a new theoretical framework for the coordinated control of two arms operating a large object. This framework not only provides a clear control of the target motion but also clarifies the contact position between the two arms and the target through control of the rolling motion upon contact. The control algorithm is a nonlinear feedback that eliminates dynamics and decouples the output. These studies on coordinated control of two arms are all aimed at specific operational requirements. Research on the operating force or torque of the two arms mainly focuses on the target operation algorithm and optimized control force. Some Japanese scholars have used fuzzy neural networks to design trajectory control and force magnitude control for the target, proposing a Delta-Bar-Delta learning ratio adaptation method for the main neural network controller. Other Japanese scholars, in their research on the manipulation and control of flexible thin plates by dual-arm manipulators, first derived the relationship between the static deformation of the thin plate and the bending stress applied to it using the Lagrange equation based on the finite element model of the thin plate. They then designed a control algorithm that manipulates the plate's motion by applying a resultant force to the plate and controls its deformation by applying internal forces. Experiments with industrial robots have proven the effectiveness of the proposed control system. Researchers in Japan, the United States, and other countries, supported by advanced manufacturing technologies and substantial research funding, are focusing on theoretical and experimental research in dual-arm robot research, particularly on coordinated control of micro-robot dual-arm movements, coordinated control of flexible dual-arm movements, and coordinated control of dual-arm manipulation of flexible objects, aiming to expand the robot's capabilities and application areas. 4. Current Status and Development in China Research on dual-arm robots in China has only been conducted for two or three years. Due to the significant gap between China's manufacturing automation technology and industrial modernization levels and those of foreign countries, research on single-arm robots is still in the initial stages of imitation and application. Therefore, research on dual-arm robots is still in its infancy. Constrained by many related technologies and research conditions, current domestic research on dual-arm robots mainly involves motion trajectory planning, dynamics, and coordinated control. Research on motion trajectory planning primarily focuses on determining collision-free path planning and coordinated motion during dual-arm operation. The Robotics Institute of Shanghai Jiao Tong University has conducted in-depth research on the time-optimal trajectory planning problem for dual-arm robots, successfully applying dynamic programming to perform time-optimal trajectory planning for the left and right arms of a dual-arm robot moving along a specific path, thus ensuring that the robot's left and right arms achieve time-optimal motion without collisions. Researchers at Harbin Institute of Technology have conducted research on autonomous planning motion control using a dual-arm free-flying space robot as a case study. By establishing a kinematic and dynamic model of the dual-arm free-flying space robot, they obtained the generalized Jacobian matrix of the robot under microgravity to describe the relationship between the end-effector velocity and the angular velocities of each joint. Then, they established a task planning algorithm and path planning algorithm for the dual-arm robot to capture targets in a floating state, and verified the correctness of the theory and the feasibility of the algorithm through a simulation system. Researchers at Xinyang Normal University, taking full account of the characteristics of dual-arm robot mechanisms, applied a recursive algorithm to establish velocity constraint equations for coordinated motion of dual-arm robot mechanisms to improve numerical calculation speed and achieve real-time control of dual-arm robots. Research on dynamics and coordinated control mainly focuses on determining the transmission performance and force transmission performance of dual-arm robots in a specified configuration along a designated direction. Researchers at the National University of Defense Technology proposed the concepts of velocity operability measure and force operability measure to guide the determination of the optimal operating posture of dual-arm robots. Researchers at Tianjin University proposed the concept, mechanical description, and properties of force constraints for controlled multibody systems. By introducing deflection (angular) velocity and generalized force, they established dynamic equations for controlled multibody systems with force constraints based on the Kane equations and achieved dynamic solutions. Researchers at Yanshan University discussed the internal force-free grasping and corresponding rhythmic load distribution methods for multi-manipulator cooperative systems for further research on the coordinated control of multi-manipulator cooperative operating systems. 5. A few views for research of dual-arm robot: Regarding the current state of research on dual- arm robots both domestically and internationally, research on dual-arm robots is still in its initial stage. Most studies are based on the condition of low-degree-of-freedom dual manipulators or two single-arm robots working together, and the research on dual-arm coordination control is mostly theoretical exploration for a specific working condition, lacking a comprehensive consideration of the entire system from the perspective of hardware and software. Therefore, as the automated machine that is closest to imitating human movements, the following points are proposed for future research on dual-arm robots: (1) As a comprehensive system composed of multiple disciplines, dual-arm robots should use a systematic approach to seek the best matching methods and implementation means for the mechanical system, control system, perception system, electronic and electrical system, etc., that make up the dual-arm robot, in order to meet the needs of theoretical research. (2) The trajectory planning and pose control of dual-arm robots should be studied using dynamic programming methods to study the motion characteristics and dynamic characteristics in three-dimensional space, especially the collision avoidance problem. (3) The research on dual-arm coordinated control and its algorithm should be based on high-degree-of-freedom conditions, and more research should be conducted on its coordinated control mechanism, real-time control, feedback means and methods. (4) Flexible operation is a major characteristic of dual-arm robot operation. Therefore, for the flexible control of dual-arm robots, especially the flexible control under redundant conditions, how to find the optimal control for the target from multiple or countless solutions. (5) How to effectively convert the target image obtained by camera into the information required for target operation, including its processing power, speed and acquisition accuracy. (6) The influence of the characteristics of the drive source on the operating force or torque of the dual-arm robot, and how to optimize the electronic and electrical circuits to meet the force or torque characteristics requirements of the dual-arm robot during operation.