Task 8 pursues three major objectives:
1) build a functional prototype of a decision support system that integrates the computational contributions of previous tasks;
2) conduct a pilot demonstration recurring to one channel for positive traffic conditioning, such as intelligent traffic light control or road message board display; and
3) evaluate the actionability of the models learned in the context of Tasks 4 to 7 in view of the goals and metrics defined in Task 2.
The target decision support system should satisfy the functional requirements to be defined in the cross-sectional Task 2, and guarantee the following quality requirements:
1) adaptability and extensibility of functionalities;
2) reliability of results by conveying guarantees of statistical significance according to the pursued principles throughout Tasks 4-6;
3) data protection and integrity;
4) persistence, proper display and share of algorithmic outputs; and
5) accessibility to algorithms recurring to web interfaces and standard APIs.
The extensible and modular development of the target system is essential to realistically account for the possibility of its deployment in other world cities.
At this stage, and in order to test new assumptions with the designed prototype, a set of new sensors for traffic analysis can be designed and placed, for a proof of concept analysis. These new sensors are expected to mitigate the challenges potentially found during Tasks 4 to 6 concerning data quality and completion, and will be used for proof of concept. The achieved preliminary results will lead to guidelines for a second phase of this project.
Finally, this task will assess the actionability of the contributions proposed along Tasks 4 to 7 in accordance with the performance indicators specified in Task 2. To this end, predictions will be assess against new incoming data. In addition, a pilot demonstration will be conducted in close cooperation with CML and other entities to evaluate the ability to positively condition traffic flow in a realistic (yet controlled) environment.
INESC-ID will be responsible for the integration of the proposed computational contributions; LNEC will be responsible for the system evaluation; and CML, closely aided by INESC-ID and LNEC, will be primarily responsible for the pilot demonstration.