| This thesis considers the real-time scheduling issues for control based applications with both timing constraints and precedence constraints in open environments. The schedulability issues of the control based real-time applications is also addressed. In this thesis, we present the real-time scheduling problem with two main goals. Our first goal is to develop a feedback scheduling algorithm to handle timing and precedence constraints in real-time environments. Our second goal is to address the schedulability issues of the control based real-time applications, where a system that cannot be scheduled with the traditional real-time scheduling algorithms can be scheduled with our proposed method.; A task-oriented scheduling algorithm (TOA) and a multi task-oriented algorithm are addressed which can satisfy both the timing constraints and the task interdependencies between/among tasks. An experimental study is conducted to assess the performance of the presented algorithms and compares them to other existing real-time scheduling algorithms. An integrated feedback scheduling model of real-time control applications is given and a formalized framework is addressed within which scheduling algorithms can be chosen based on timing constraints and desired quality of service.; We develop a multi-robot scheduling architecture, an offline scheduling algorithm, and an online scheduling algorithm that provide performance guarantees in terms of deadline miss ratio and CPU utilization in face of unpredictable task arrivals and execution time variations due to network latency and jitters. These scheduling algorithms are analytically designed and tuned with feedback control theory. For a given set of tasks with bounded time periods, tasks that are not schedulable of their given frequencies may be schedulable with the proposed feedback scheduling method. |
