HAPADS concept, shown in Fig 1, will be based on a mobile platform mounted on Vehicles for data collecting, processing, modelling and transmitting wirelessly. As the wireless transmission of large data from mobile device cannot be available, there is a need of processing data locally. The proposed platform will integrate the sensors and transmitting devices, it will also extensively collect and process data. Thanks to the flexibility of the programmable circuits and systems, the platform can be developed together with data processing algorithms, providing flexible matching the processing power to the requirements of the algorithms.  To monitor gas such as NO2, HAPADS method is based on the dedicated nitrogen dioxide gas sensors developed with the utilization of metal oxides that serves as gas-sensitive materials. The measurement method will be based on the novel microwave circuits and conventionally resistive measurements. The main advantage of the HAPADS is that the limit of detection for target gases will be adjusted for this specific application, to cover the full range of possible target gases concentration in comparison with commercially available NO2 detectors.  To monitor PMs, HAPADS methods are using direct particle counting of certain sizes allow precise and unambiguous measurements. The proposed sensor uses this principle to accurately measure PM. Currently there are no portable sensors available on the market that use this measurement principle. The innovativeness of the proposed solution lies in the development of the first prototype using the principle of direct PM counting, which is small, cheap and can be used as a portable device. The key component in the proposed PM sensor is an image sensor with high sensitivity and low noise.

Fig. 1 HAPADS concept

HAPADS development is implemented in four main work packages (WPs1-4) and one supporting WP (WP5):

  • WP1 - Detector development for MPs (AGH, GUT, NILU). WP1 objective is to develop new detectors of PM and NO2 detection where the limit of detection as well as detection range can be adjusted for the specific conditions, being overcomes the challenges of mobile sensing solutions due to moving velocity and wide range conditions of the outdoor weather. When sensor mounted on vehicles with the limitation of the system moving velocity, only a short integration time is allowed, which will result in a low signal-to-noise ratio (SNR) in dark illumination conditions. Therefore, time-delay-integration (TDI) image sensors can be applied to solve the problem.  Moreover, the gas sensor should be developed in the technology that can be easily combined with CMOS technology to design and develop a full product, a gas sensor with read-out circuits; a lab-on-chip module.
  • WP2 – Programmable multiprocessor hardware for data acquisition and signal processing with parallel edge computing and deep learning algorithms support (GUT, NILU). The WP2 objective is to devise new specialized hardware for MPs that suitable for mobile sensing solution and edge intelligence. This WP  will develop a new automatic data acquisition system (DAS) based on programmable devices and accelerators (e.g., Field Programmable Gate Arrays (FPGA), Intel Movidus NCS, Google edge TPU), to overcome the growing problem of existing mobile DASs on acquiring and analysing data at the required sampling rate. In addition, This WP will devise the optimization of a mobile platform concerning performance and robustness of radio connectivity.
  • WP3 - Multi-objective optimization of MPs (UiT, PWR, NILU). The new embedded software will make each MP autonomous and cognitive, being able to adapt its policy for an optimal balance between multiple objectives such as sharp accuracy and energy efficiency. The optimal policy enables MPs to self-calibrate for mobile deployment.
  • WP4 - Development of advanced models for improving accuracy of local MPs operating within dynamic environments (NILU, GUT, UiT). WP4 objective is the development of correction layer for improving MP measurements accuracy which will use data-driven models and machine learning (ML). The correction models will combine differential data exchanged between multiple sensors with information on local environmental conditions. Once fed with data, the models will address measurements inaccuracy resulting from dynamics of the environment.
  • WP5 - Integration, Validation, Testing and Data Services for Drivers and Publics (PWR, GUT, NILU, LIG) - the supporting Work Package.