Periodic Reporting for period 2 - InDrive (Automotive EGNSS Receiver for High Integrity Applications on the Drive)
Reporting period: 2017-01-01 to 2017-12-31
The excellence which the proposal pursues is sustained by the following items.
• Technological excellence.
o The idea of heterogeneous positioning gets extended, leading to the definition of a generalized close coupling, leveraging data fusion and including E-GNSS-related information, on-board sensors, map-matching, and also information coming from cameras and from the connected vehicles.
o The concept of integrity is extensively investigated and addressed also when the heterogeneous positioning is concerned – not only for pure E-GNSS positioning.
o The ADAS solution built in InDRIVE will rely on such integrity information, usually neglected in vehicular applications.
• Scientific plane subtenduing novel theoretical investigations:
o the use of Bayesian tool as a mean for the fusion of very heterogeneous information;
o the study of new signal observations of GNSSes and the evaluation of their potential effect on the accuracy and integrity of the position; notably this analysis will leverage a custom real-time software-receiver available among the partners.
• InDRIVE will foster close-to-market solutions. The Consortium includes Partners that have already practically tested automatic driving; the automation Level 2-3 which InDRIVE pursues is indeed close to market.
• The project also aims at unlocking a strong market potential: while the proposers do not exclude to integrate also expensive sensors (such as LIDARs), especially for higher levels of driving automation, in this proposal they primarily investigate the beneficial role of less expensive ones. The proposed architecture is meant as a mass-market solution.
• The inDRIVE solution will also be beneficial to the hot topic of Vulnerable Road Users (VRUs) safety. This constitutes indeed an open issue in the current state of the art: pedestrians and cyclists will hardly wear or be equipped with a VANET. InDrive proposes an alternative solution coping with this.
• The activities of the Q1 have mainly dealt with the initial set-up of the project, with the use cases and therevision of the architecture.
- Use cases are defined: each partner has proposed some use cases; they have been jointly discussed; the most significant ones are those which highlight benefits of InDrive solution. The shortlist has been brainstormed through a questionnaire which has been distributed internally and externally.
- Compared with what was defined in the proposal, the refined architecture has taken into account all the feasibility and optimization aspects which the Partners identified. The architecture has been further (slightly) refined during the project, to deal with more implementation criticalities.
• The activities of the Q2 mainly dealt with the final use-case definition, with further refinement of the architecture, and the initial implementation of the multipath algorithms.
- After collecting feedback from the stakeholders, the use cases have been shortlisted, ranked and selected.
- The architecture has been further investigated. The implementation/porting of SW-receiver has been started; studies on the Bayesian approach have been carried-out.
- Dissemination: website launched; initial links to other projects; participation in the EU ITS congress; acceptance of ITS WC paper; ideas for some additional papers
• At the end of the Q3, the target platformshave been selected and most of them purchased. This has let emerge also the on-board integration issues and led to some internal modulation of expenses. Additionally:
- Procedures for early testing are put in place, as a risk countermeasure
- Business analysis started
• Also the activities of theQ4 well reflect the plans of the GA:
- The implementation of the GNSS components has proceeded. Some preliminary tests carried out.
- The “Probabilistic Map Matching” has also made progresses. Integration of lane detections mostly completed. Progresses and partial results on all the enabling technologies - partially demonstrated at the mid-term
- The deployment of the in-vehicle application and of the integration have been kicked-off
- The preliminary business case and plan is refined
From the perspective of positioning, this project is to develop a key enabling technology to leverage EGNSS for automated manoeuvres and safety critical decisions in automotive applications. This will be done by a combination of low level probabilistic signal processing algorithms within a software defined GNSS receiver, in combination with advanced high level data fusion approaches, in order to derive a confidence estimate that is able to meet any safety requirements in urban areas.
InDrive condenses probabilistic GNSS signal tracking algorithms for software defined receivers and gives an extension towards the explicit integration of multiple received signals. This will be done by explicitly utilizing the benefits of Galileo signal in space properties.
2) FRAMEWORK OF AUTOMATED DRIVING
The National Highway Traffic Safety Administration (NHTSA) has defined five levels of vehicle automation. InDrive will work with L2 and L3 automation. Main expected progresses from the project are the level of accuracy in positioning in urban areas that can be obtained with mass market receivers and the level of accuracy in the understanding of the obtained level of accuracy. In fact, for commercial use of ADAS and autonomous systems in urban scenarios it is important to obtain a good level of accuracy in position with low cost receivers that can be adopted for commercial applications and to know the level of accuracy in position that can be achieved.
3) GNSS SOFTWARE RECEIVER
The InDrive architecture proposes the use of a software receiver to feed the Bayesian filter with all the necessary information in real-time, exploiting the intrinsic flexibility offered by the software architecture.
In modern communications, the flexibility is a key point in the implementation of receivers because it eases the development of new products and the introduction of new features in already existing ones. In the silicon-based products the extra cost associated to the introduction of a new functionality is mainly due to the development process rather than the extra silicon and components. On the other hand, full reconfigurability can be obtained far more cheaply through a software based implementation: the well-known Software-Defined-Radio (SDR) approach.
4) VANETS FOR COOPERATIVE DRIVING FOR VRUS’ SAFETY
It has already been demonstrated that it is possible that a car can even drive autonomously on its own, with no connections to a controller, or to the other vehicle . On the other hand, the availability of a VANET network might simplify the work and widen the impact - thus it would be a waste not to use them.
InDrive proposes to use VANET for:
• Including all the information detected by InDrive architecture within the electronic horizon.
• For VRUs: the information can be generated on board and not by any future transceivers customized for VRUs. This means using VANETs for VRUs y rather than creating new VANETs to cope with VRUs.