For many years, dynamic modeling of robotic arms has been a well-known method in academia. It is commonly used as a tool in robotics and engineering laboratories for developing simulation robots and advanced control algorithms, motion simulation, and other academic pursuits. Dynamic modeling involves the development of mathematical formulas to describe the robot's dynamic characteristics, such as inertia, mass, center of gravity, and other properties that are not easily measured. Although frequently used in theoretical research, the application of dynamic modeling methods to improve robot control has long been overlooked by industrial robot developers and manufacturers.
Servotronix first considered the potential of dynamics modeling to address the performance challenges experienced by a manufacturer of high-speed Delta robots used in the semiconductor industry.
Applications involving the use of robots in solar cell processing demand high acceleration performance and extremely high precision. However, due to its inherently brittle structure, the Delta robot is very prone to breakage. Furthermore, it poses a threat of impact and damage to expensive production loads and materials.
The Delta robot poses a risk of damage to itself and its payload.
In some systems, the incremental motion of a robotic arm is based on a mechanical parallelogram connected to a moving platform and arm linkage via ball joints. If certain positions or angles are exceeded, the force required to disintegrate the robotic arm will be significantly reduced, making it highly susceptible to disintegration in the event of a collision or forceful push or pull. Further complicating this situation is that these break points are often located at extended positions, posing a significant risk of impact with obstacles. If the impact consequences of the disintegration go undetected, further damage may result.
To overcome these shortcomings and better control the Delta robot, Servotronix engineers adopted and enhanced a dynamic model, which originated from academic research.
The resulting model, or the set of algorithms describing the Delta manipulator, is only half the story.
This is because the model is generalizable. However, since robots have different sizes and masses, the actual parameter values will vary. Although the kinematic principles of robots are the same, the physical characteristics of each robot are different. Even within a series of robots that have been manufactured, the physical characteristics of each robot may differ slightly, which will have different effects on performance.
Once the robotic system has been modeled, accurate values for its dynamic parameters are needed. Servotronix achieves this by developing additional algorithms to automatically determine the kinematic and dynamic parameters.
While some parameters, such as the geometry of the robot arm links, can be easily measured and applied to formulas, other parameter values, such as the center of gravity of each link, are determined by an automatic identification program.
During the recognition process, the robot will move randomly, and the values of its dynamic parameters are determined using Servotronix's recognition algorithm. Both internal and external factors are incorporated into the calculations, such as shape, material, cables, and friction.
Following the successful implementation of model-controlled Delta robots, Servotronix has further developed dynamic models for other robot types, such as SCARA robots, 4-axis cross-arm robots, and 5-axis Galilean spherical robots. The company recognizes that dynamic modeling can meet the growing demand for higher output, faster speeds, and lower costs in robotics.
Establish a dynamic model for a 5-axis Galilean spherical robot
By using dynamic models, Servotronix customers have achieved faster settling times and better motion trajectory control for their robots. This approach also has the added benefit of allowing the detection of system wear through changes in mechanical parameters, particularly the friction constant, over time.
As the torque error demonstrates, the calculated torque value almost accurately predicts the transmission torque value.
Model-based control is now a built-in feature of the Servotronix softMC multi-axis controller. Its effectiveness is ultimately achieved through a real-time motion bus system. EtherCAT allows the Servotronix softMC multi-axis controller to update drive values every millisecond. During each sampling period, the softMC sends commands to and receives feedback torque values, as well as standard position and velocity values, from the drive. The softMC is typically used in conjunction with Servotronix CDHD servo drives. The softMC can be configured to treat the received torque as an additional value, supplementing the self-calculated torque values.
Servotronix softMC3 and softMC7 (from left to right) multi-axis motion controllers employing model control and real-time EtherCAT motion bus.
The direct benefit of model-based control, as demonstrated in the Delta robot case, is the ability to detect and avoid collisions, thereby providing better protection for the load, work area, and operator. Furthermore, it eliminates the need for force sensors, simplifying system design and reducing costs.
The most significant advantage of this control method is the enhanced robot's behavioral capabilities and actuation performance. Because the torque required to reach a given position can be calculated and controlled with greater precision, the robot's motion path can be highly optimized. The required current is also more stable, as the current data can be calculated rather than simply obtained through a feedback loop, thus providing better speed control and reducing the possibility of the robotic arm trembling or sudden movements.
By assessing the torque and force required for the robot's movement and preventing excessive torque, the robot's speed can be increased safely and easily, while reducing oscillation and settling times. The end result is that model-based control allows the system to move faster and more accurately, achieving higher processing capabilities.
