PH‐RLS: A parallel hybrid recursive least square algorithm for self‐mixing interferometric laser sensor

The authors present the parallel-hybrid recursive least square (PH-RLS) algorithm for an accurate self-mixing interferometric laser vibration sensor coupled with accelerometer under industrial conditions. Previously, this was achieved by using a conventional RLS to cancel parasitic vibrations where itself is not in stationary environment. This operates sequential mode and due its compute data-intensive nature, does work real-time applications, hence requires parallel computing. Therefore, existing C program parallelized hybrid OpenACe C/MPI (Open Accelerators/Message Passing Interface) programming models tested on Barcelona Supercomputing Center CTE-Power9 Supercomputer. computational performance of proposed PH-RLS compared code executing multi distributed processors uni-core processor architecture, respectively. While comparing eight nodes supercomputer, results show that gets 5857 times improvement as implementation single node system. also gives scalable different range signals, making it suitable choice interferometer sensing systems working