The paradigm of autonomous machines has shifted with the remarkable advancement in machine intelligence. To support machine intelligence, autonomous machines are now equipped with diverse sensors, heterogeneous multicore processors, and distributed computing nodes that require complex software architecture to utilize them properly. With the introduction of new sensors and computing powers, autonomous machines must now support applications that performs complex processing on unbounded sequences of stream data produced at real time. However, with the increase in software complexity it is becoming difficult for developers to coordinate the multiple streams of data and still meet the system requirements. To tackle such difficulty, we are currently developing a graphical programming framework we named Splash. Splash provides effective programming abstractions that allow the users to establish multiple stream processing applications effortlessly. Splash also gives users the ability to specify genuine end to end timing constraints required by their system. The timing constraints in turn are automatically monitored for their violations by the Splash framework. This thesis will introduce the components of the Splash graphical programming framework and focus on how Splash provides stream processing capabilities for its applications. The thesis will also introduce the internal workings of Splash’s monitoring capability for end-to-end system timing constraints. Lastly the thesis will validate Splash application’s functional correctness and tests its timing constraint monitoring capability by implementing ACC (Adaptive Cruise Control) and LKAS (Lane Keeping Assistance System) algorithm using Splash.
【 预 览 】
附件列表
Files
Size
Format
View
Designing and Implementing Core Runtime Libraries for Splash Programming Framework