Medical sensors once embedded in the human body are hard to test for their selectivity to analyse (glucose, uric acid, etc.), response time, accuracy & precision, sensitivity and reproducibility of results. The maximum detectable limit of analytes and their quantification w.r.t. changes in analyte concentration vary among sensors. The stochastic behavior of biosensors (and their environment) call for their probabilistic modelling and verification of important characteristics like expected response time, linear quantifiable limit, etc. DGB Technologies has successfully verified important measures of interest of a bioreceptor (in a biosensor) for one of its clients Integrated MedicalSensors, developing artificial pancreas for humans, using state-of-the-art model checker STORM.
DGB has employed route optimization algorithms for a number of clients in the transportation sector. Our solutions find the shortest possible route, given a road network of locations, from source to destination that optimizes cost and time. In addition to the employed algorithms, we have also worked closely with a North American client to help them in the verification of an existing employed route optimization solution in the Consumer Goods delivery sector using Formal Methods techniques. Our solutions have resulted in a significant increase in revenues by adopting route optimization solutions.
Industries (e.g., petrochemical plants) are increasingly using safety instrumented systems (SIS) to complement their process control systems in order to reduce the risks of accidents. SIS is composed of sensors, logic solvers, and final control elements that take a process to a safe state when predetermined conditions are violated. The components of SIS, therefore, include pressure and temperature sensors, a control system, control valves or other final control devices, electrical wiring, process piping, power supplies, software, etc. The probability of failure of SIS on demand (PFD) -- which is based on its individual components -- defines the risk that is acceptable if an accident occurs. IEC 61508 has quantized acceptable risks into four levels, called Safety Integrity Level (SIL), that can be assigned to SIS -- SIL 4 SIS is considered to be the safest. DGB Technologies successfully completed SIL analysis of a "Fire alarm & sprinkler system for skyscrapers" using dynamic fault trees (DFTs) and STORM model-checker. We proved that by figuring out sensitive and non-sensitive (whose duplication minutely improves the overall SIL level) components, a cost-effective and reliable system can be designed.
DGB has worked closely with a number of higher education institutions across the world to build early prediction warning systems for students at-risk of dropping out. Our Long short-term memory (LSTM) based predictive model seamlessly integrates on top of existing learning management systems of clients to help reduce dropout students and also to identify features that can guide decision makers to help improve extrinsic motivation factors in online learning setup. Our solutions have helped a number of clients to increase their students' attendance along with their learning outcomes to further meet their institutional KPIs.
DGB is building a framework for a US client to automate financial anomaly detection in the power sector financial transactions. The anomaly detection framework can effectively help in catching the fraud, discovering strange activity in large and sophisticated Big Data sets. Our solution searches and increase the effectiveness of our client’s digital business initiatives. Using Machine Learning approaches on a big data platform, the framework distinguishes between the spike detection and point of change detection - which has helped reduce our client’s losses significantly.