Collaborative software integrity verification using bit inversion error checking codes in crowd-coupled iot applications

Internet-of-Things (IoT)-based pedestrian crowd activity monitoring is now being applied in our lives. The application as a service, that is stored in on-chip flash memory in an IoT sensor processor, performs collaborative software execution between the connected IoT systems and the monitored human activity. The software code memory integrity of embedded flash memory in the IoT microcontroller (MCU) is becoming important for applications requiring safety-critical operations because the latest services directly affect human activity. Software-driven or hardware supports for the memory protection are required to guarantee the safe-conscious code execution of the downloaded firmware in MCUs. The protection method requires more power consumption in decoding the additional padding bits. In this paper, the protection hardware and software technique in the IoT systems are proposed to improve the safety and integrity of the software execution with a small amount of logic gates overhead. The error correction code (ECC) hardware data path is integrated with our newly designed binary bit-inversion decoder with zero overhead inversion flags to decrease the power consumption in decoding the binary code blocks. The 8-bit ECC per 64-bit code blocks (72:64 SEC-DED) is applied to a 64KB flash memory with the separated region of the flash memory for ECC padding bits. The proposed robust software execution unit is successfully integrated with human crowd monitoring system in the library.