An Automated Guided Vehicle (AGV) begins to move in a warehouse. It sets off, accelerates, avoids an obstacle, and reaches its destination shortly afterwards. To an observer, this is nothing remarkable. The vehicle simply drives. Yet that is rarely as straightforward as it appears – and that is why hardly any two of these movements are exactly alike. Before the vehicle even begins to move, its journey has al- ready been defined. A modern logistics center generates trans- port orders, prioritizes flows of goods, and calculates routes. Navigation often relies on software frameworks such as the Robot Operating System 2 (ROS 2). These specify not only where the vehicle should go, but also how it should get there. The planning process produces a motion profile containing position, velocity, and acceleration. This planned motion is known as a trajectory. It is precise and fully calculated – and it works, at least on screen. Turning a trajectory into motion For the vehicle to actually move, this description must be translated into action. Motion control begins where a set- point – for position, velocity, or current, for example – is con- verted into controlled, dynamically stable motion. The motion controller receives the setpoints from the nav- igation system and translates them into specific control signals for the drive. A geometric description becomes a time-depen- dent reference signal that must match the dynamics of the entire system. This information is often transmitted via stan- dardized interfaces such as CANopen. A path through space thus becomes target values for speed, position, or torque. Inside the drive, a process begins that continues throughout the entire journey. The controller operates continuously, com- paring the specified setpoints with the measured actual values. Encoders capture the motor’s current motion, while sensors provide additional information. Based on the deviation, the controller calculates the re- quired motor response for the next instant. It responds not only to the current deviation, but also takes into account how the system will behave next. This process runs continuously, many thousands of times per second. Motion is not generated all at once, but adjusted continuously. From setpoint to torque At the end of this control chain is an electrical value. The controller specifies a current setpoint, which the power elec- tronics apply to the motor phases. In the brushless DC motor, this creates a rotating magnetic field that interacts with the rotor. The resulting torque at the motor shaft is transferred to the wheel through a gearhead. In the calculated model, the vehicle follows the trajectory exactly. Reality is different. A wheel briefly loses traction, the load changes or the surface varies. The measured motion then no longer fully matches the planned motion. The vehicle be- gins to deviate from its intended trajectory. What matters is not the deviation itself, but how quickly and reliably the system responds. The controller detects and corrects the deviation in real time, keeping the motion stable. Where it gets complex Each individual step is understandable on its own. In practice, however, they are closely interconnected. Navigation, trajec-
tory planning, closed-loop control, and physical execution take place simultaneously and influence one another. What happens within the system is not a linear sequence, but an interplay. Why there is no off-the-shelf solution This is where it becomes clear why such systems cannot sim- ply be transferred from one application to another. An AGV in a logistics center behaves differently from a mobile robot in a production environment. The differences result from weight, dynamics, surroundings, and precision requirements. As a result, the requirements placed on the drive also change.
A path through space becomes target values for speed, position or torque.
The motion calculated at the beginning must be tailored to that specific application. This concerns not only individual parameters, but the design of the entire system – from closed- loop control to mechanics. The mechatronic behavior of the system must therefore be understood in detail. Only this un- derstanding makes it possible to tailor the drivetrain and con- trol system specifically to the application. In the end, the vehicle completes its route. From the out- side, the process appears simple. Inside, the motion has been planned, translated, executed and continuously corrected.
About the author Fabian Vogel is Managing Director of maxon zub machine control AG.
Together with his team, he develops motion control solutions for complex mechatronic drive systems and supports customers in implementing demanding automation and robotics applications.
47
Motion Control
Made with FlippingBook Publishing Software