Theoretical Framework for Distributed Biomedical Signal Processing in Edge Computing Architectures for Nigerian Rural Telehealth Networks

Telehealth networks in rural Nigerian settings face fundamental constraints of intermittent and low-bandwidth connectivity that render cloud-dependent biomedical signal processing architectures unreliable, motivating the development of edge computing architectures that can maintain clinical signal processing utility under connectivity-impaired conditions. This dissertation develops an original theoretical framework for distributed biomedical signal processing in edge computing architectures specifically designed for rural Nigerian telehealth network conditions. The theoretical framework addresses three interrelated engineering challenges whose intersection defines the unique problem space: the heterogeneous and non-stationary connectivity environment of rural Nigerian telehealth networks; the real-time processing requirements of clinical physiological signals including ECG, EEG, and multi-parameter vital sign streams; and the hardware resource constraints of edge computing nodes deployable within the cost and power consumption limits appropriate for rural Nigerian health post settings. Original theoretical contributions include a distributed signal processing partition theory specifying how biomedical signal processing algorithms can be decomposed into latency-critical and latency-tolerant computation stages and dynamically allocated between edge and cloud based on available connectivity, a communication-aware quality of service theory for biomedical signal streaming that formalises the clinical signal quality degradation consequences of compression and transmission error under varying connectivity conditions, and an edge node resource allocation theory for concurrent multi-signal processing under battery power constraints. The framework is instantiated as a reference architecture and validated through simulation using empirical connectivity traces from six rural telehealth sites in Taraba and Kebbi states and through prototype implementation on Raspberry Pi-based edge nodes at two sites. Keywords: edge computing, biomedical signal processing, distributed architecture, rural telehealth, Nigeria.