Accuracy of Sensor-Less Control of an Electro-Mechanical Brake 2015-01-2704
At first glance a friction brake should be controlled by normal force to produce predictable brake force. Controlling an actuator (and hence brake pad) position basically seems to introduce uncertainties to normal force and brake force, because at first view the position tells little about actual normal force.
The electro-mechanical brake (EMB) of Vienna Engineering (VE) can be operated by position-control, either without force sensor (saving costs) or even with a true brake torque sensor. For position control a relation between actuator position and normal force is used. When pad wear is correctly adjusted a certain actuator position produces a given deformation and at known elasticity the deformation produces a defined normal force. In the VE-EMB this relation uses a three-dimensional curve and includes temperature influence of the coefficient of friction, thermal expansion and thermal elasticity change.
The temperatures for this control can't directly be measured at normal vehicles, but an accurate model that can be easily calculated in a low speed processor loop at the brake controller is possible.
Position control showed to be an useful control method for the VE-EMB and even has advantages to force-control (as e.g. used at conventional brakes): when actuation losses in a force controlled system increase then the real normal force is lower than expected, reducing the accuracy of brake force. With a position controlled brake the actuation always goes to the same motor angle, always producing the same normal force in the brake.