Filed while leading SaaS / IoT telematics at Bosch — not in a lab. Each one took raw connected-vehicle data and an AI/ML model and made it predict something the field couldn’t see coming: a failing pump, a tiring battery, a missing sensor. Two are granted; all four sit on platforms that shipped, scaled and stayed.
Early-warning prediction of fuel-injection-pump failure using telematics data and a machine-learning model — fuel level, engine speed, coolant temperature and hydraulic pressure streamed device-to-cloud. The model that anchored a 95%-accuracy predictive-maintenance platform.
An intelligent, data-driven model that estimates a battery’s State of Health and Remaining Useful Life from an unknown state — no prior battery history required. Built for battery-electric vehicles and industrial / home appliances running on Li-ion cells.
In ESP active-safety systems the master-cylinder pressure sensor is crucial — and often absent on low-end programmes. This replaces the physical sensor (or its rough estimation model) with an AI-trained model that delivers accurate estimated pressure, including as a fallback on sensor failure.
A big-data framework over connected-vehicle telematics — cloud, distributed storage and advanced data mining — analysing use cases from predicting reaction time and per-kilometre trip cost to CO₂ emission, vehicle-health diagnosis and driver-behaviour classification.
Patent counts are vanity unless they ship. These four are interesting only because the telematics platforms that carried them shipped, scaled, and stayed.