Print This Page Print This Page

  • InDrive Project
  • Our Scientific Approach
  • Target Scenarios
  • Use Cases

InDrive Project

The main objective of InDrive project is to develop and demonstrate innovative close-to-market applications, which are heavily relying on accurate and high integrity satellite navigation. To achieve the full potential of advanced satellite positioning, an integrated solution starting from low-level signal processing to high-level data fusion will be proposed to get a continuous probabilistic positioning of high integrity.

InDrive will demonstrate the future use of mass-market GNSS, targeting automotive applications with high demands for integrity by creating a framework that specifies the requirements for data acquisition, signal tracking and data fusion in order to guarantee the proper handling of positioning data. This approach introduces an innovative integrity framework, allowing the applications to comply with their specified false alarm rates.


The innovation of this project is to leverage EGNSS localization for automated manoeuvres in automotive applications. In order to meet the requirements for each of these use cases we will introduce a technology for innovative confidence computation. By implementing this approach we can guarantee the compliance of the use cases in terms of false alarm rates and accuracy.  The idea of applications with different demands on integrity is shown in the following image. Within the red area, safety critical applications, as emergency breaking will operate. The extension to the yellow area will meet the requirements for warning based applications with lower time to collision demands. The green area represents the informative applications, like traffic or weather information.

Our Scientific Approach

The general idea of InDrive project is the development of an automotive enhanced positioning platform, based on the integration of GNSS and other on-board sensors that is able to meet the requirements of safety critical applications in the context of autonomous driving and driver assistance. The core is to adopt existing technologies for the improvement in terms of availability and accuracy like EGNOS/EDAS, on-board inertial and odometry information and to complement these approaches by an integral Bayesian data fusion environment for a consistent monitoring of the integrity. This framework will integrate a software defined GNSS receiver in order to implement and adopt innovative signal tracking algorithms for improving the integrity of the solution even under worst environmental influences like Non-Line-of-Sight and Multipath errors in urban areas. The holistic Bayesian approach, starting from the receiver antenna until the evaluation of the estimated position and the related confidence within the final application will ensure the integrity of the decisions based on the provided position estimate.

There is a range of adaptive use cases that will profit from the dynamic integrity concept. In the course of the project an evaluation of the requirements of relevant use cases will be given. A simple use case for the dynamic integrity concept and fusion of position uncertainty and application is given in the Figure 1 and Figure 2. The driving instructions of the given scenario can be adjusted according to the current uncertainty of the position estimate which is represented by the orange cloud. As this example is not directly safety relevant, but the user acceptance will be dramatically improved as no wrong driving instructions will be given to the driver. The next evolution of this concept is the deduction of automated manoeuvres based on this information. In a cooperative GNSS based ACC or Platooning the adoption of the integrity concept will result in a dynamic adaption of the distance between subsequent vehicles in order to follow the regulations of safety distances between vehicles.


Figure 1 – Position confidence is high, driving instructions can be very precise. “Please switch the lane and turn into the second road.”



Figure 2 – Position confidence is low and ambiguous, driving instructions need to be soft. “Please be sure to be on the right lane and prepare turn right. One of the next roads is yours.”

To guarantee a consistent confidence computation, it is required to implement a probabilistic approach in each of the signal processing steps, starting from the receiver antenna until the position on the map. We will propose a positioning framework that integrates the work from previous projects on GNSS low-level signal processing with special attention on the GALILEO signal properties and the work from high-level application based projects with the focus on data fusion and integration.

  •      Proposed architecture for E-GNSS receiver

This idea is complemented by an innovative multipath integration strategy by integrated probabilistic modelling of ambiguities on signal level. Multiple received signals will not be resolved on signal level, but fused within the Bayesian framework which propagates the uncertainties of these local influences into the final position confidence. This helps to map the current integrity of the system to the demands of the current application and will enable the usage of EGNSS for safety critical automotive applications.

The handling of all uncertainties coming from environmental influences and their propagation into the application is handled by a holistic Bayesian framework which is based on a modular approach and is thus able to improve the position estimate with a range of augmentation technologies. Here, InDrive will adopt most relevant strategies in order to provide the best possible availability and accuracy. Figure 3 shows the central Bayesian Framework component that integrates all technologies to be adopted for the positioning filter. The interfaces from additional sensors to the positioning filter will be derived from the Bayesian filtering theory. All sensors are integrated by their measurement models and sensor models. This approach guarantees a flexible and modular approach in order to easily adapt the system to different environments and requirements. The following sensors and technologies will be integrated:

  • GNSS Integration: Provides the interface to the InDrive GNSS receiver and is thus the integral part of the filter.
  • Vision Integration: Provides additional knowledge regarding road and lane conditions in order to improve the lateral positioning performance.
  • Odometry Integration: Provides wheel speed and yaw rate of the vehicle and helps the system to improve the availability and accuracy under GNSS signal outage.
  • EGNOS/EDAS: Provides information about the ionospherical influences and satellite health and timing parameters in order to improve the accuracy.
  • Map Integration: Position information will be brought into the context of a map. The map is included also as sensor, as the system estimates the position on the map and the regarding lane statistically.
  • Other Sensors: The system provides an open architecture and is thus open for any extension from other external sources like LIDAR/RADAR or additional inertial sensors.


Figure 3: The Positioning Filter is designed as a modular system, which foresees sensors and extensions that will be implemented within the project and is open for a flexible integration of other requirements and sensors.

  •             The role of a software-GNSS receiver

The majority of consumer GNSS receivers are essentially black boxes that output a stream of messages containing position, speed, heading and other information. There is an unknown amount of processing done to smooth the output signal and reduce the noise, the properties of which are generally considered proprietary and information about post-processing is not made available. In any case, this additional processing step varies between manufacturers and models. This “gray area” about the processing inside the receiver is not acceptable in the context of the Bayesian navigation filter introduced above, because it needs far more flexibility in accessing the low-level observables within the signal tracking stages.

In order to fulfil this need, the InDrive architecture proposes the use of a real-time software receiver to feed the Bayesian filter with all the necessary information, exploiting the intrinsic flexibility offered by the software architecture.

The intended SDR platform will be beneficial to the overall InDrive solution in terms of:

  • Availability of low-level signal statistics for inferring the integrity of the GNSS positioning;
    • virtually, any observable signal attribute might be made available to the Bayesian framework;
    • this could lead to the identification of new relevant observables.
  • Real-time availability of the observables: this may be indeed a differentiator. In fact,  in principle, a delayed and averaged signal evaluation may lead to less precise estimations; while an additional punctual and real-time analysis might lead to more precise determinations.
  • Possibility to operate on some of the core signal processing algorithms of (A1) to refine their control through a dedicated SW interface with (A2).

The algorithms implemented in a versatile SW-receiver can be modified and tailored to the target application, and any required observable can be easily made accessible and available to the successive system stages.

InDrive will demonstrate several applications in the area of Advanced Driver Assistance Systems (ADAS) and future Intelligent Transportation Systems (ITS) based on different positioning requirements. We will consider connected and not connected vehicles:

  • Use case «vehicles that are not connected» target is to measure the performances of integrity and accuracy of the position so to «go on the map» but always with mass market receivers. The map section that is needed, we can download temporarily and locally in the car. In this use case we insert the proposed new algorithms for the receiver. This «positioning» can effectively serve vehicles that are already equipped with sensors for ADAS – major task here in «to do the positioning».
  • Use case «connected vehicles» major task here is «to reconstruct the scenario around the vehicle» in here we can develop the evolutionary concept from information systems on safety to integrated 360 deg. ADAS functions to automated manoeuvres targeting level of automation 3.

It is clear that both a more robust and precise positioning and the awareness on the accuracy will be two factors that will influence the automatic manoeuvres. Considering these factors, InDrive use cases are addressed based on a 3-step incremental approach:

  • Before leaving

This step involves all operations that are needed and can be performed before or while starting the trip and does not necessarily require the vehicle to be “connected”. The provided vehicle is equipped with a navigation system that includes e-horizon component and digital maps, all additional information that is envisaged as needed for the forthcoming trip – such as ADAS attributes for the map section that will be used – can be downloaded and temporarily stored locally on board (for instance, at home and then transferred to the on board system through a Smartphone or other personal devices, or directly by the on board system if equipped with connectivity). This step relies also on the availability of position information whose reliability and accuracy are measured and validated; thus, specific Test Cases will be designed to assess the positioning system (mass market receivers + innovative algorithms) performances according to some predefined KPIs.

  • Approaching the critical areas

Future trajectories can be predicted within the single vehicle by the e-horizon. This step requires the vehicle to be “connected”: through connectivity, these trajectories can be exchanged between automated and not automated vehicles, for instance by means of periodic beaconing messages. Similarly, other data on intended manoeuvres can be shared, including driver’s or co-driver’s future intentions. These data may span from simple future “turn” manoeuvre indications to more complex speed and acceleration profiles. In addition, connectivity among automated vehicles can enable more advance schemes such as manoeuvre negotiation, the formation of coordinated and co-operative groups of vehicles and so on. However, reliability of data has to be properly handled by the recipient vehicle unit, considering two main aspects: a common map referencing among cooperative vehicles, and position accuracy.

  • Operating manoeuvres

Finally, the benefits deriving from improved positioning, map update, and use of V2X connectivity can be evaluated on automated driving scenarios. This activity will be performed on demonstrator vehicles where the performances obtained with the prototypes developed in the project can be compared with the previous standard solutions considered as a reference. Here particular attention will be given to urban situations where the positioning suffers from more critical situations and it is important to know the intention of other vehicles in situations such as crossings or roundabouts.

Several use cases will be considered for evaluating the benefits deriving from the adoption of the integrity concept into the automotive industry. In particular, they will be selected with a particular emphasis on Advanced Driver Assistance Systems (ADAS) and autonomous driving solutions with the aim of identifying how to benefit from the position confidence information in such a context.

Three different application areas will be taken into account: Safety, Comfort, and Green driving.  By doing that it will be possible to broaden the scope of the project and improve the market acceptance of the identified solutions.