Technology
Home > Technology
Context-aware Model-based AI:
Superior AI technology for E-Mobility
Fleet Digital Twin - AI for Behavior Learning
The WideSense AI IoT platform automatically learns high-precision predictive Digital Twins containing behavior models of vehicle charging, fleet drivers’ driving styles, fleet vehicles and vehicle components.
Adapted from the research of UC Berkeley's Model Predictive Controls Lab, these dynamically composed merged models can be learned in a fraction of the time of purely statistical methods. Obviating the need for GPUs and specialized AI processors.
Our patented technology employs data-efficient and highly scalable model learning, resulting in a compact representation of fleet Digital Twins. Behavior model compactness further enables real-time capability and the ability to continuously feed predictions to WideSense Fleet Optimization & Recommendation engines that can solve multi-objective optimization problems with context-sensitive constraints in near real-time.
Planning and Scheduling
The introduction of EVs has created a new need to limit the assignment of certain vehicles to certain trips, particularly due to the limited and significant variability of range of EVs.
Our platform builds optimized efficient schedules over month horizons that satisfy demand, maximize service, minimize operator cost, and meet the budgeted operator headcount.
Vehicle Real-time Work Assignment
Limited and highly variable range and long refueling (charge) times necessitate careful assignment of work to vehicle trips. The WideSense platform selects work for each EV based on the energy demand of the work and the capacity, layout and queues of the charging network. It ensures vehicle availability for pull-out, avoids blockages, minimizes confusion, and avoids vehicle swaps (critical for EVs due to their limited range). Schedules are produced daily and are updated in real-time as the day progresses and conditions change.
Real Time Operations Support
Throughout the day, WideSense predictions are combined with real time information on vehicle and charging infrastructure status, weather and traffic patterns, as well as road conditions. Daily vehicle use plans are compared to these predictions and real time alerts of plan changes are dispatched via email, text and the Smart EV Operations Management portal to the teams responsible for successful fleet operations. Changes to work assignment and charging schedules are presented in short, simple guidance.
Maintenance Insights
The WideSense AI Digital Twins automatically learn in actual driving conditions the State of Health, including the performance, aging and degradation of each vehicle’s key components such as battery, drivetrain, braking and regen. Each component’s performance can be compared historically, to extract trends of performance degradation or spot sudden changes. In addition, the component performance can be compared to similar vehicles in the same fleet to find low-performing outliers. The Digital Twins also provides accurate forecasts of the component Remaining Useful Life (RUL) that can inform maintenance scheduling to reduce unanticipated vehicle downtime.
Fleet Optimization & Recommendations
The WideSense AI Cloud platform delivers predictions of energy demand, battery charge time and vehicle component State of Health. These predictions support planning , scheduling, operations and maintenance of vehicles and charge infrastructure. The platform allows fleet owners to maximize fleet operations performance including maximizing vehicle & charger utilization and minimizing energy costs within their unique operational constraints.
AI Cloud
WideSense Digital Twin learning, prediction and fleet operations optimization all take place in the cloud, connecting contextual resources such as weather, traffic and routing with each fleet’s information management configuration. Integration support is available with fleet CAD/AVL, trip planning, scheduling and telematics systems and is based on public standards where available (GTFS) . Fleet owners can connect their existing IT tools to the WideSense AI Cloud and get fleet insights and real-time guidance within days.
Cloud hosting relieves the fleet owner from any IT management complexities and shortens the time to implementation. Cloud hosting allows the on-demand scaling of the WideSense AI system as the EV fleet grows both in size and complexity with vehicles having varying specifications including battery capacity.
IT Security
WideSense protects its application and customer data with a layered approach.
Data is classified and sensitive data is segregated from the app platform. All external and maintenance connections to the app platform are encrypted and protected via user authentication and authorization. User rights are based on roles and organization. They are checked and validated on every request to application services.
All business data is managed by Tier-1 cloud service providers. Backup infrastructure is available for each core business IT service. Application infrastructure is provided by distributed, hosted service providers like Amazon’s AWS, Microsoft Cloud, or Google Cloud. Each application is packaged in a quickly deployable container that can be moved between industry-standard base OS instances. All data and processing servers are remotely hosted at physically secured data centers at Amazon’s AWS or Microsoft Cloud or Google Cloud.
Data and network security is provided by industry-standard firewalled clusters, always encrypted data connections outside of clusters, pub/priv key-based authentication to system assets, and password rotation policies.
Ready to get started?
