BlossomEdge AI

Description

BlossomEdge has developed innovative technology to detect hazards and other relevant events in micromobility, such as (e-)bike, cargo (e-)bikes, e-scooters, mopeds, and motorcycles.

The core are AI models using computer vision and various sensor inputs (e.g. gyroscope, accelerometer, GPS, magnetometer…) to perceive the world “through the eyes of a rider”. The models are hardware-agnostic and run in real-time on the edge, even on conventional smartphones. They also can be integrated into existing applications or hardware stacks and fused with existing data.

The technology is already used by hundreds of users in >20 cities worldwide. This computer vision and fleet telematics technology creates meaningful risk and safety insights at various abstraction layers. These include insights about individual situations or entire rides, individual vehicles or entire fleets, individual locations, streets, or cities.

–

Applications include:

• Rider risk or safety scoring / quantification

• Infrastructure evaluation and planning input for risk / safety potential

• Hazard perception as realtime input for autonomous vehicles

• Routing input to determine safe / risky routes

• Rider training and feedback to change behaviour

• Various research applications in the realm of objective and perceived safety

BlossomEdge is partnering with universities, cities, and companies to find market applications and drive R&D, and are actively looking for public funding projects.

–

Example research topics include (but open to explore others):

• Target-group specific risk perception (e.g. perceiving risk on a bike like a child, an elderly person, a beginner…)

• International or city-by-city comparison of rider (risk) behavior, incl. subgroups such as delivery riders, youths on e-scooters

• International or city-by-city comparison of country / city risk profiles

• Effectiveness of infrastructure measures in terms of perceived and objective risk

• Integration of better use of data and technology solutions to enable effective multimodality