An Experimental Study on the Embedded Software Development Approach Using TI-RTOS

Abstract

Embedded systems have experienced significant growth in recent years. These systems, which power a wide range of devices from smart appliances to industrial machinery, have become integral to our daily lives and industries. However, this rapid expansion introduces new challenges for developers. A critical decision involves whether to employ a Real-Time Operating System (RTOS) or to forgo it, depending on the specific project requirements. Additionally, developers are focused on finding effective strategies to enhance code quality, promote code reuse, reduce complexity, and streamline the learning curve. These challenges underscore the evolving landscape of embedded systems, where maximizing potential while maintaining efficiency and ease of development has become a top priority. This study presents a comparative evaluation of an embedded system utilizing the TI-RTOS, which includes an RTOS kernel and a set of libraries, against an equivalent system using a minimalistic custom kernel without these libraries. The analysis examines various metrics, including source code size, complexity, architecture, learning curve, and development time. The results indicate that the learning curve associated with adopting TI-RTOS did not show a significant increase compared to the system without an operating system (bare metal). Furthermore, productivity remained largely unchanged when using TI-RTOS. Importantly, the implementation of TI-RTOS did not lead to a notable increase in source code complexity. This study offers valuable insights for embedded systems developers and engineers, demonstrating that integrating TI-RTOS can be a viable option without adding undue complexity to a project. These findings are particularly relevant for those seeking an efficient and user-friendly solution for embedded systems.