Time optimal path control for industrial robots

Industrial robots are used in repeated tasks. The path followed by the robot differs with the change of product. In mass production line product change frequencies are longer then a month. In order to increase the efficiency of the robot, task duration must be decreased. Because of this reason robot motions must be planned.Robot motion is not only used for efficiency but also used for task specific requirements. Robot motion planning is the most significant part of robot control at industrial robot systems. The motion have to be done by satisfying given conditions. Defined path have to be tracked in terms of given jerk, acceleration, velocity and position for each actuator of the robot. Optimal robot motion planning has a significant role when maximizing the level of efficiency of the robot system. Dynamic and kinematic optimization is improved by optimizing the robot path according to the criteria for fully utilizing its capacity and take full performance of the actuator. In industrial applications, last ten years it is easy to get high process capacity computer for controlling the robot. As the effectiveness of the optimization methods increases, computer-aided motion planning are widely used day by day. Robot producers are focused on the optimal motion planning. The mechanical properties of the same class robots are almost same. Major difference that can be added to robot is done by newer software. Computer-aided robot motion planning, optimization reduces the work load on the operator which increases the production capacity of the robot while reducing downtime. For one cycle optimal motion planning can reduce about miliseconds or seconds. These miliseconds or seconds can increase the time efficiency of the motion about fifty percent. In industrial applications such as welding, bonding and painting robot repeats the same motion at the mass production line. Trajectory planning application should be made as fast as possible or as precise as possible. In applications common used criterias are minimum energy, minumum time and minimum jerk. Predetermined path have to be optimized under certain restrictions according to the specific tasks in the optimization problems. SOA spatial operator algebra methods for kinematic and dynamic equations are used. Generally Denavit Hartenberg method is use to solve kinematic and dynamic equations of robot. SOA method is an advance method which is very easy when it needs to update kinematic and dynamic equations of motion planning. This method has simple stucture when used in varying degrees of freedom. It is widely used when solving complex robot mechanisms. Optimal robot motion planning is based on optimizing the robot actuator. It bases on time, energy and vibration criteria in practice while taking into account the limits of the actuators. Optimization saturates at least one actuator of the robot in a sample time of the optimization. It also includes the calculation according to the initial and last states in the application or by the desired optimization of the trajectory in operational space. The motion planning problem is classified by the optimization criteria. The optimization method used in this thesis is named time-dependent optimal trajectory planning which minimizes the time. If the optimization criteria is energy, it will be named as energy dependent optimal trajectory planning. This thesis is based on the general motion planning problems as well as the optimization, especially considering the time-dependent optimal robot motion planning and time-energy-dependent optimal robot motion planning problem. The optimization criteria in the optimal trajectory planning have to be inclueded in the generalized equation. Objective function is defined in order to minimize or maximize the value of optimization criteria. System limits such as speed, acceleration, jerk and torque are defined in the optimization problem. Industrial robot arm has to complete the motion without surplusing the speed, acceleration, jerk and torque limits of the actuators. Besides optimization criteria in the optimal trajectory planning speed, acceleration and torque constraints have significance. The determination of the path can be reached in two ways: Kinematic and dynamic trajectory planning trajectory planning. Kinematic trajectory planning does not deal with values mass, inertia, force and torque. In Dynamic trajectory planning values; mass, inertia, force and torque are taken into account. Obstacle avoidance can be applied for optimization by using genetic algorithms. In order to complete path in minimum time or minimum energy adaptive genetic algorithms, the second order cone programming, particle swarm optimization techniques can be used. In the trajectory optimization firstly optimization problem is converted to convex optimization problem beacuse of solving optimization Convex optimization problem is generalized in terms of notation used in this thesis. Generalized optimization problem is solved by using second order cone programming method. The optimization purposes that used in this thesis can be summarized as completing the path in minimum time. Circular trajectory tracking of robot and linear trajectory tracking applications are investigated in simulation. Trajectory of the robot at circular path, time is reduced from 3.6 seconds to 1.5 seconds after the optimization method applied. While trajectory time decreases the energy of the actuator increases. At circular path pseudo energy that indicates mean value of actuators torques is increased 5 times higher than the energy before the optimization. In addition at the linear path tracking time of the robot is reduced from 4.0 seconds to 1.2 seconds. It can concluded from these results time optimality of the robot motion planning is successfully achieved. The software is tested in the offline simulation environment later in this study are likely to be tested on a industrial robot arm. Simulation results indicates that the solution gathered in the study have reasonable values. Real system control will be a bit different. The time spended for the optimization calculation can be reduced. In a subsequent study of the optimal trajectory planning of robot will be examined online and is intended to work to accelerate the optimization algorithm for it. Online time optimal control or online energy optimal control will be very useful when operating to the industrial robot arm.