An embedded system architecture for acoustic signal capturing and orientation

For the light weight acoustic reconnaissance applications, an embedded architecture design including hardware and software is proposed in this paper. The novel architecture model is based on dual-DSP in which master DSP cooperates with slave DSP to speed up the data processing while keeping low power consumption. This paper especially focuses on the hard and software implementations of this model. By using HPI and daisy chain techniques, communication and emulator interface is simplified in the hardware implementation. In the software implementation, interrupt techniques and embedded Linux kernel are adopted and plane 4-element cross array orientation algorithm is transplanted to our dual-DSP model. Experiment results indicate that this architecture design is successful and can be used actually for acoustic reconnaissance applications.