Actors: Inria / ACCENTAURI, Cerema / GIPI, Inria / FUN, Inria / COATI, Cerema / ENDSUM, Cerema / GeoCoD, Cerema / DTerSO, Cerema / DTerMed.
Data collection is at the heart of the integrated management of road infrastructure and OA. However, these captures are by definition carried out in very restrictive environments (eg: reduced accessibility), even hostile (eg: weather conditions). The sensors are all of different natures with heterogeneous dimensions, sensitivity, means of communication, providing heterogeneous data in size, type and acquisition frequencies. Collecting this data therefore presents a real challenge, which must involve agile and adaptive communications protocols, deployments of autonomous robot fleets, service vehicle route planning as well as the implementation of new services such as geolocation. The data collected under this axis 1 will feed into axes 2 (modeling) and 3 (knowledge extraction). Conversely, he will receive recommendations from these two axes to improve the precision and relevance of the data.
This axis is split into 3 tasks:
Task 1.1: Sensor deployment study
Actors: Inria / FUN, Inria / COATI, Inria / ACENTAURI, Cerema / ENDSUM, Cerema / DTerSO, Cerema / DTerMed.
This task will study for different use cases the most suitable capture locations for each data. This study will take into account business needs but also physical deployment constraints (accessibility, radio environment, possibility of electrical supply or ambient energy recovery) and deployment needs for collection. It will recommend the most suitable means to collect the data according to the places of collection, the necessary frequency and the costs of collection: radio collection, multi-hops, visit of robots / drones, etc.
Task 1.2: Geolocation services
Actors: Inria / ACENTAURI, Inria / FUN, Cerema / ENDSUM, Cerema / DTerSO, Cerema / DTerMed, Cerema / GéoCoD.
In order to be able to take measurements reliably and position them precisely, on-board geolocation services are necessary. They can be used for the navigation of drones or autonomous robots, for trajectory planning or even to locate the location of the capture. Each application will bring its own implementation constraints and precision requirements, therefore calling on specific and varied positioning methods. The different techniques that will be used, adapted and combined for each study include among others radiolocation, SLAM and GPS. This task will aim to design a positioning technique to meet the application needs in order to provide the other tasks of this WP with precise location information.
Task 1.3: Self-deployment of robot fleet
Actors: Inria / ACENTAURI, Inria / FUN, Cerema / ENDSUM, Cerema / DTerSO, Cerema / DTerMed, Cerema / GéoCoD, Cerema / GITEX.
In some cases, the collection of information may require the dispatch of a fleet of autonomous robots, on the ground or in the air (drones). The goal of this task is to design the self-deployment techniques to allow each entity in the fleet to know where to move, how to move, while maintaining connection with others and performing the requested task. These techniques and different mechanisms will be local and adaptive algorithms which will take as input the application constraints and the place in which the data must be collected (via pipes, high up under bridges, at different points of a dam, etc.). They will be combined with the results of task T1.2 (localization) and potentially multi-hop communication protocols as defined in task T1.1. These methods will be accompanied by assistance in piloting the drone: automation of certain piloting functions (for example stabilizing at a given distance from a structure, maintaining a constant speed) and situation analysis for a transition to manual piloting in the event of of problems.
Task 1.4: Mission planning and dynamic adaptation
Actors: Inria / COATI, Inria / ACCENTAURI, Cerema / GIPI.
We are interested in optimizing vehicle routing to measure the condition of the roads. The initial planning of a vehicle’s route seeks to minimize the time required to travel through all of the routes to be analyzed. This planning can take into account various constraints such as the need to pass several times on a wide road, fuel consumption or even the driver’s break times. However, measurement errors may occur when crossing certain roads (the trajectory of the vehicle is not perfect) and it is then necessary to return to these roads to acquire the missing data.