Self-Healing System on Chip based on runtime reconfigurable Field-Programmable Gate Array (FPGA)


Technology abstract

An Italian SME working in the Space field is offering a reconfigurable self-healing integrated system (ReSHeTS). ReSHeTS is based on a System-on-Chip (SoC) embedding a COTS Reconfigurable FPGA able to guarantee operational flexibility (reconfigurable according to the mission objective) and radiation tolerance (reconfigurable in case of Single-Event-Effects). This system can be used as the main core of an On-Board Computer (OBC), enhancing both reliability and flexibility while reducing the cost.

Technology Description

ReSHeTS is based on a System-on-Chip (SoC) composed of a microprocessor, a partially run-time reconfigurable COTS FPGA, and a rad-hard memory. Such configuration exploits the hardware acceleration of FPGA, that represents the main computing unit, and the higher performance of COTS components with respect to Space qualified FPGA. A rad-hard microprocessor with low performance is required to monitor the FPGA and to check if any error occurs. Taking advantage of FPGA reconfigurability, the System on Chip is able to counteract possible errors induced for example by radiation. In fact, if any error is detected by the microprocessor, the microprocessor will overwrite the damaged module of the FPGA with a golden copy stored in a rad-hard memory. In such way it is possible to automatically restore the SOC functionality without requiring external intervention or replacement of components. The ReSHeTS architectures can be used also to exploit the same computing power to perform different tasks; in fact, it is possible to automatically change the FPGA modules dedicated to the execution of different operations.

Potential applications of this technology span from the aerospace field, where reliability is a fundamental requirement, to all the industrial fields where high-performance computing and the ability of changing the implemented functionality can be used to speed-up data elaboration. Moreover, combining reliability and adaptability, ReSHeTS can be used to realize safe and powerful computing cores while reducing the dimensions. This can be exploited in biomedical devices, as well as for on-board computer in the automotive industry.

A prototype of the System on Chip has been already developed and tested. The system has been developed with the goal of reliability and radiation tolerance; thus, it has been tested by exploiting a fault injection technique which emulates Single Event Effects.

Innovations & Advantages

The described technology increases system flexibility as well as reliability, even in case of single event effects. Moreover, it exploits COTS devices, which have a number of advantages with respect to traditional rad-hard components. The main advantages can be summarised in the following points:
• for devices with equivalent performances, COTS components have lower cost (one or more orders of magnitude);
• for devices with comparable price, COTS components have higher performances.
Therefore, being able to exploit commercial components is an interesting solution. ReSHeTS permits to greatly reduce the limitations of the devices that shall operate in a radiative environment, allowing to use COTS components in deep space missions or in environments with a high level of radiation.
Another advantage of this technology is represented by its flexibility. An on-board computer implementing the ReSHeTS architecture can easily reconfigure itself in order to perform new elaborations on data. For instance, if the telecommunication protocol between ground station and the satellite is upgraded, the OBC can be adapted just by modifying its decoder logic. This ability allows to extend the mission lifetime, as well as increase the computational power of the satellite.

Further Information

ReSheTS has been tested exploiting a fault injection technique which emulates in a realistic way the effects of radiations on electronic devices. One million errors have been injected into the FPGA. The system has been able to correct all of them, except for 0.0036% of cases in which the faults hit internal communication interfaces, causing a processor stop. However, this issue can be easily solved exploiting a watchdog system.

Current and Potential Domains of Application

Aerospace field
Automotive field
All industrial fields where radiation tolerance or high-performance computing and the ability of changing the implemented functionality can be used to speed-up data elaboration.