Control System Basics
A control system (also called a controller) manages a system’s operation so that the system’s response approximates commanded behavior. A common example of a control system is the cruise control in an automobile: The cruise control manipulates the throttle setting so that the vehicle speed tracks the commanded speed provided by the driver.
In years past, mechanical or electrical hardware components performed most control functions in technological systems. When hardware solutions were insufficient, continuous human participation in the control loop was necessary.
In modern system designs, embedded processors have taken over many control functions. A well-designed embedded controller can provide excellent system performance under widely varying operating conditions. To ensure a consistently high level of performance and robustness, an embedded control system must be carefully designed and thoroughly tested.
Integer Algorithms: Implementation and Issues
Low-end embedded processors generally lack hardware support for floating-point arithmetic. Although software floating-point implementations are sometimes available in library form, the execution time of the routines tends to be prohibitive. Floating-point representations also consume substantially more data memory than the common integer data types.
Because of these drawbacks, developers tend to avoid floating-point arithmetic in low-end embedded applications and implement mathematical operations using integer arithmetic. While the speed and memory advantages of integer arithmetic over floating-point are substantial, it is important to exercise the utmost care in designing integer algorithms. For those coming from the PC development environment (where the use of floating-point is ubiquitous), significant a conceptual adjustment may be necessary. This article highlights some integer arithmetic rules of thumb and demonstrates integer algorithm implementation in the context of an example PID cruise control application.
Simulation Planning
Before developing a simulation, a plan must be created identifying what is to be simulated, the degree of simulation fidelity required, and how the resulting output data will be used. Many additional details must be addressed as well. There must be a simulation software development process that deals with issues such as configuration management, release criteria, and problem reporting and resolution. There must also be a manageable process that identifies how the simulation is to be operated, including the definition of input data sets, and how simulation output data is analyzed and used as input for decision making. To achieve maximum value from the simulation effort, the simulation plan must be thoroughly integrated into the larger project plan.
