Heterogeneity in distributed embedded systems

Meer over het boek

Advances in wireless sensor networks (WSNs) have resulted in a plethora of hardware platforms and operating systems for embedded devices. The emergence of the Internet of Things (IoT) has further increased this diversity, leading to a range of embedded systems from microcontrollers to smartphones. However, this variety has not been adequately addressed by any universal software architecture. While multiple operating systems exist, they typically support only a limited number of architectures. For instance, Java cannot run on tiny microcontrollers due to the absence of a Java Virtual Machine (JVM). Additionally, testing environments are predominantly homogeneous, making it challenging to evaluate concepts across large, heterogeneous networks. This work introduces techniques to tackle these challenges by applying modern programming paradigms from desktop computing to embedded systems. We developed a testing platform featuring heterogeneous wireless nodes and sensors. The dominant programming language remains C, but we demonstrate that more efficient results can be achieved with C++ by utilizing template-based designs from libraries like the Standard Template Library (STL) or Boost. Our implementation, Wiselib, serves as a generic algorithms library for distributed heterogeneous embedded systems. We constructed a testbed with various nodes and sensors, including self-designed load sensors for tracking passers-by, connected to se