Research On A Multi-Linked And Tracked Rescue Robot And Its Motion Strategy

Rescue robots have been applied broadly in nature disasters (such as flood, fire, and earthquake et al.) and various man-made catastrophes. At present, various disasters happen frequently and seriously endanger folks' life and safety. After an accident happens, it is difficult for rescuers to make salvation at the first time in the abominable site. To acquire exact information of the location of trapped or dead people, temperature, content of oxygen and deleterious gas and collapse status in the site, it has an important practical signification to develop a rescue robot who can substitute rescuers wholly or partially to go to the disaster field for environment detecting.Based on an analysis for the surroundings in the disaster site, according to the function requirement of a rescue robot, a design rule is made. In addition, the mechanism of a multi-linked rescue robot with tracks is presented and developed, which has stronger motion ability. A tracked moving mechanism is used to improve the motion capacity of the robot in complex environment. Each link is enwrapped by tracks, which not only increases the drive force, but also improves the ability of turning and avoiding being locked. A motor is used to drive the ambient tracks. Transmission mechanism uses a bevel gear chain mechanism, so not only the transmission chain is short, but also the structure is compact and the tracks lay symmetrically. Each joint mechanism has a pitch and a yaw degree of freedom, whose motion ranges are±45o. The positive pressure anti-explosion technology is adopted to design the reliability of the rescue robot when there is a lot of flammable gas in the disaster site.A forward kinematics model of the rescue robot is built for the posture adjustment motion of the head unit. Meanwhile, its inverse kinematics is mainly studied. A new space decomposition method is presented to decompose the joint space and working space, and a hybrid algorithm based on the combination of space decomposition and genetic algorithm is presented to resolve the problem that the inverse kinematics has not analytic solutions. This algorithm effectively reduces the searching space of genetic algorithm and improves the solution precision. A dynamics model of the rescue robot based on Newton-Euler method is built for its posture adjustment motion of the head unit. Furthermore, the static and dynamic stability is analyzed. The stability criterion presented supplies a theory foundation for realizing stable posture adjustment. The action theory between the robot and a terrain is analyzed and a dynamics model by iterative Newton-Euler method is built for the rescue robot's movement.An unknown environment modeling method based on a visual tangent graph is presented and its relative terms are defined and described too. The visual tangent graph notion is expatiated and extended. A path-planning algorithm is proposed based on the visual tangent graph. The algorithm uses the visual tangent graph to describe the unknown environment and combines the reactive path planning and the reflective path planning to plan a path from an original point to a destination. The simulation analysis indicates not only the visual tangent graph is an effective path planning tool, but also the path planning algorithm based on the visual tangent graph is feasible, which increases the environment adaptability of the rescue robot, reduces the motion time and satisfies the motion requirement of the rescue robot in unknown environment. To follow the certain optimum path for the rescue robot, a following kinematical model based on Follow-the-leader strategy is built and the two grate nested optimization of genetic algorithm is presented to search the optimum dynamical solutions.The autonomous obstacle negotiating motion function of the rescue robot is built. The typical obstacle negotiating ability of the rescue robot is analysed and the obstacle negotiating motion is discussed based on the key posture kinematics and dynamics. According to the rescue robot's mechanism and the character of environment, an autonomous obstacle negotiating layered planning algorithm based on an expert control system is proposed. The autonomous obstacle negotiating algorithm is divided into behavior planning and motion planning. An off-line planning of each obstacle negotiating behavior and relative motion is made to form a rule base. A real-time on-line reasoning based on behavior inference engine and motion inference engine is carried to realize the antonomous obstacle negotiating function of the rescue robot.An experiment system of the rescue robot is established. The experiment on the basic motion such as linear and turning movement is made. Meanwhile, the experiments on the path planning algorithm based on the visual tangent graph and autonomous obstacle negotiating planning strategy based on the expert control system are done. The results show that the rescue robot proposed has stronger motion ability and the path planning algorithm and autonomous obstacle negotiating planning strategy proposed are correct.