Robot control can be quite challenging because there are so many factors to iron out with sensing, coordination, and overall system reliability. Robots need to tackle various situations and, at any point in time, may experience mechanical or electrical failures. Most of the major issues, though, involve information obtained from sensors, which may sometimes contain noisy or incorrect data. Some of the best tips to incorporate in robot control include filtering out inaccurate sensor data, including time-outs, and having a task-control-based programming structure.
Unreliable sensor data is one of the most common issues that cripple a robot's functioning. For example, a sensor reading may go undetected, or it may give an inaccurate result. It may become corrupted because of external conditions. If this unreliable data is sent to the high-level routines, then the robot is hampered in its task. To combat such a situation, it would be better to include specific routines during the programming stage that check the data and pass it on to high-level routines.
For example, a proximity sensor could feed data in to an object detection routine. If the sensor gets jammed and it keeps sending in an unreasonable distance for the object's location, the routine should be able to figure this out. Not including a data check at this stage results in a wrong figure being passed on to higher robot control processes. The robot may then just grind to a halt because it can't figure out if there's an object in front of it or not. So, putting in routines that check the data for specific sensors before they are sent off to be interpreted helps in smoother robot functioning in robot control.
Including time-outs in the robot control processes is also beneficial because it helps the robot to exit an endless loop in case something goes wrong. For instance, if the robot is supposed to go forward, but the collision sensors are faulty, it may keep hitting a wall in front of it repeatedly. Unless there is some time-based exit condition incorporated into the program, it could remain stuck in the same task forever. It is helpful to include a maximum time limit for a task within the programming routines. If the robot fails to execute the task within the specified time, then the time-out helps it to exit the loop.
It's also beneficial to incorporate task-oriented robot control within the programming to break up the steps. For instance, a robot's task may be to pick an object, turn left 90 degrees, and place the object down. In this case, the program can treat the entire process as a single task with initial and final conditions, an activity that has to be performed, and a return value. A sensor-specific function such as "object_pick()" could return a value of true or not depending upon whether the robot picked the object. This approach helps the robot to function smoothly and makes it easier to identify problem areas.
