Bridging Real-Time Robotics and Computer Architecture

  Robotics technology, poised to play a pivotal role in shaping future societal frameworks, is projected to reach a  deployment of 20 million units and a market valuation of US$70 billion by the end of this decade. Despite its growing  significance, there remains a pronounced gap between the fields of robotics and computer architecture research. This  thesis endeavors to bridge this gap, contributing to the advancement of real-time robotic systems through two primary  initiatives: (i) the development of comprehensive, open-source benchmark suites coupled with systematic performance  studies, and (ii) the devising of efficient computer architectures for robotics.  

Firstly, the thesis introduces extensive benchmark suites for robotics tasks and applications. These suites are  accompanied by thorough performance analyses, delving into the computational behaviors and architectural implica tions of the applications. Such suites and performance studies are instrumental in deepening the understanding and  improving the synergy between robotics and computer architecture.  

Secondly, the thesis proposes efficient computer architectures for robotics. This involves the development of  application-specific hardware accelerators, designed to optimize crucial robotic kernels such as collision detection, and  the development of domain-specific processors. These processors are not exclusively tailored for particular kernels  but are optimized to deliver enhanced performance across a wider range of robotic applications. The empirical results  demonstrate significant speedup potentials, achieving up to 41.4 times with application-specific hardware accelerators  and 3.8 times with domain-specific processors.  

Collectively, these contributions not only propel the performance capabilities of real-time robotic systems but also  establish a groundwork for ongoing and future research at the critical nexus of robotics and computer architecture.