Loading...
Loading...
Internet of Things, sensor networks, and embedded computing
6 Problem Statements in this Domain (5 Standard + 1 Open Innovation)
India has a large stock of aging bridges, many in remote locations without reliable grid power near the structure, and structural health monitoring (detecting developing cracks, abnormal vibration, deck deflection) currently depends on periodic manual inspection that can miss slow-developing structural degradation between inspection cycles. A continuous monitoring solution needs to operate for years on harvested or minimal battery power while still capturing the meaningful structural signals (vibration modes, strain patterns) that indicate developing problems — a genuine power-budget-versus-sensing-fidelity engineering trade-off.
Rural piped water schemes under the Jal Jeevan Mission need to verify that water is actually flowing reliably and safely to household connections, but existing monitoring approaches assume stable power, reliable connectivity, and controlled installation conditions rarely available in remote rural distribution networks — where pipes are frequently tampered with, sensors face fouling from sediment-laden water, and communication infrastructure is patchy. The challenge is maintaining trustworthy flow and basic quality (e.g., turbidity, contamination indicators) monitoring despite this genuinely hostile deployment environment.
Distribution transformers serving rural feeders fail frequently due to overloading, poor maintenance, and extreme heat, and failures cause extended outages for entire villages because replacement transformers and skilled technicians are not always close by — yet most rural transformers have no monitoring at all, and retrofitting expensive commercial monitoring systems across the enormous number of rural transformers is not economically viable. A workable solution must detect developing faults (winding degradation, oil issues, overheating) early enough to enable preventive maintenance, using sensing that is dramatically cheaper than standard SCADA-integrated monitoring.
Voice-based agricultural advisory kiosks could meaningfully improve access to expert guidance for farmers with limited literacy, but real deployment environments involve significant background noise (market/field settings), regional dialect and code-mixed language variation not well covered by mainstream speech models, and unreliable or absent internet connectivity ruling out cloud-based speech processing for most interactions. The genuine embedded-systems and speech-processing challenge is running usable, dialect-robust, noise-robust speech understanding largely on-device with constrained compute.
Many Indian cities have old, poorly-mapped underground water and gas distribution networks with significant unaccounted-for losses from leaks, but comprehensive instrumentation (pressure/flow sensors at every segment) is prohibitively expensive to retrofit across an entire legacy network, and pipe location/condition records are often incomplete or inaccurate. Effective leak detection and rough localization must work with sparse instrumentation — reasoning about the network as a whole from a limited number of strategically placed sensors rather than assuming dense coverage.
Have your own innovation in IoT & Embedded Hardware? Architect your own smart sensor telemetry network, edge AI device, industrial IoT monitoring hardware, or intelligent asset tracker utilizing microcontrollers, low-power radios, and real-time cloud sync.