Design and Optimization of a RISC-V Architecture Based on the Chiplet Paradigm for Nanosatellite Applications
Embedded Systems; RISC-V; Chiplets; Nanosatellites; New Space.
The growth of the New Space sector has driven the development of missions based on nanosatellites and CubeSats, increasing the demand for onboard computers capable of combining low cost, flexibility, reliability, and performance. However, most currently employed solutions are based on proprietary and monolithic architectures, whose development and manufacturing involve high costs and limited adaptability to the specific requirements of different missions. In this context, this dissertation proposes a modular computing platform for nanosatellite onboard computers based on the integration of the open RISC-V instruction set architecture and the chiplet integration paradigm. The choice of RISC-V is motivated by the elimination of licensing costs and by the possibility of customizing the architecture for space applications, while the adoption of chiplets seeks to exploit modularity, component reuse, heterogeneous integration, reduced manufacturing costs, and platform evolution. The research comprises the identification of requirements for space embedded systems, the design of the proposed architecture, its implementation in a hardware description language, the modeling of an interconnect inspired by the UCIe specification, and its validation through functional simulations and FPGA synthesis. As part of the validation, different architectural scenarios are investigated, including multicore organizations and different memory subsystem configurations, enabling the analysis of platform performance, scalability, and resource utilization. The results demonstrate the technical feasibility of integrating RISC-V and chiplets for nanosatellite applications, indicating that this approach represents a promising alternative for the development of more accessible, modular, and evolvable onboard computers, while contributing to the broader adoption of open technologies in the New Space context.