The project successfully worked on the four objectives during two cycles. Each of the cycles consisted of requirements definition, development, integration and piloting phases, concentrating on technology development in the start, and the validation of the proofs of concept in the end. The objectives were realised as follows:
O1: Development of perception sensors and low-level sensor data processing was done and included advancements in FMCW and UWB radar technology for human vital sign monitoring, Lidar for environment detection, IMU for both automotive and human motion detection and diaphoresis sensors.
O2: Data processing and ML algorithms were developed for the processing and fusion of sensor data in order to derive the features for the realisation of the use case requirements. The algorithms provided e.g. improved vital sign extraction from radar and foil-based sensors, behaviour extraction and classification from a combination of remote and motion sensors, improved accuracy of positioning of road users from IMU, camera and radar based measurements, emotion and distraction of drivers and improved estimation of exertion level.
O3: A reference architecture was developed to capture the components, functionality, and connectivity necessary to realise distributed sensing and computation. System integration exemplifying this architecture was done for each of the use cases resulting in functioning demonstrator systems.
O4: Demonstrators were developed for each of the use cases and piloted in lab or real environments. An overview of the use cases is given below:
UC1 Integral Vitality Monitoring – This use case focuses on elderly and activity monitoring and has four demonstrators validated in six pilots:
- Continuous activity monitoring - indoor positioning, radar-based vital sign monitoring, and camera assisted activity monitoring for the elderly and activity monitoring solutions piloted in the laboratory and the Lleida living lab in Spain.
- Episodic Health Gate - a radar-based vital sign monitoring and weight sensing chair for frailty detection in elderly piloted in a care facility in Kuopio, Finland.
- Wearable Activity Monitoring - a wearable devices-based exercise monitoring solution including sweat analysis piloted in outdoor exercises.
- Sleep Monitoring - a radar-based sleep monitoring solution piloted in Kempenhaeghe, The Netherlands
UC2 Driver monitoring – This use case addresses the topic of driver monitoring in the context of partial automated driving. It includes three demonstrators and two pilots:
- Driver simulator – integrated system for assessing the Driver Complex State (DCS, a combination of cognitive, behavioural, and emotional states) and derived a “fitness to drive index” by means of different unobtrusive sensors (cameras, biofeedback sensors, driving data).
- Passenger car and heavy-duty vehicle – both car and truck feature an FMCW radar to monitor the driver’s vital signs as well as a gaze detection system. Pilot validation was done in real traffic situations.
UC3 Safety and Comfort at Intersections – constitutes three main topics:
- Vehicle - has developed sensing solutions and intelligent algorithms for improving ego-localisation and vehicle environment monitoring – particularly for vulnerable road user (VRU) detection by means of IMU’s, Lidar, cameras and vehicle data.
- Communication + GNSS - worked on a UWB based positioning system with accurate synchronisation and highly secure solutions to improve GNSS based positioning robustness and reliability.
- Infrastructure - focused on the development of road-side sensing solutions utilizing radar, camera, thermal camera and time of flight camera for the detection of VRU’s and their behaviour.