One of the critical functions in autonomous driving is vehicle localisation, if it is accurate and reliable it allows the safe navigation of the vehicle when operating without human supervision (SAE Levels 4 and 5 of autonomous driving). This paper presents the performance results obtained when testing in real conditions an innovative positioning engine based on a safety-oriented paradigm. The tested engine employs a mass market dual-frequency GNSS receiver, intelligent off-the shelf automotive cameras, accurate navigation maps, low-cost inertial sensors and vehicle odometry. A unique feature is the provided integrity layer, estimated in real-time and consisting of Protection Levels (PL), which bound the error of each estimated value with a certain confidence level. This integrity layer is used as part of autonomous vehicle architecture to ensure that the vehicle navigates safely. The results show the position performances along with the real confidence levels achieved by the estimated Protection Levels. The solution estimated by the engine in real time is based on the dual-frequency measurements obtained from a mass-market GNSS receiver, the measurements of the lateral distance to the lane-markings provided by an intelligent off-the shelf automotive camera, lane-level accurate navigation maps, the measurements obtained from low-cost inertial sensors and the vehicle odometry. With respect to the engine, it is able to process all these measurements by means of: - A real time PPP hybrid algorithm that employs dual-frequency GPS and Galileo measurements, inertial sensors and PPP corrections obtained from a web server when the connection is available through the cellular network. - A robust hybrid GNSS standard positioning algorithm allowing to perform consistency checks in parallel - An algorithm that improves the accuracy by means of the vehicle sensors, the lateral distance measurements to the road lane markings provided by the automotive intelligent camera and the lane-level accurate maps. It is also able to provide an accurate position relative to the map. The position and orientation values provided by the engine (including the position computed relative to the map) are complemented with the estimation of the associated integrity Protection Levels (PL), computed for multiple target integrity risks. The implementation of an integrity layer is crucial since in safety-critical applications it can be more important to know whether the information is reliable than the precise information itself. This integrity layer determines the degree of usability of the location and orientation estimations, which is used as part of autonomous vehicle architecture to ensure that the vehicle operates safely. The position and integrity engine has been integrated in an autonomous car and has been tested in real conditions under different environments assessing its performances based on the accurate reference provided by other equipment installed in the car for testing purposes. The paper presents the position and protection level performance results obtained when testing the vehicle under open-sky, sub-urban and deep urban conditions. The engine employed to obtain the results presented in this paper has been developed and tested with an autonomous car within the ESCAPE project ([1]). ESCAPE (European Safety Critical Applications Positioning Engine) is a project co-funded under the Fundamental Elements program of the European GNSS Agency (GSA). It started on October 2016 with a duration of 3 years and with the main objective of developing a localisation system that provides the vehicle position and attitude estimations to be employed in safety critical applications like Autonomous Driving (AD) or Advanced Driving Assistance Systems (ADAS). The project is led by the Spanish company FICOSA in collaboration with partners from across Europe: Renault, IFSSTAR and the University of Technology of Compiegne ( UTC) from France, STMicroelectronics and LINKS Foundation from Italy and GMV from Spain. ESCAPE enables a high-grade of data fusion (GNSS, inertial sensors, cameras and vehicle sensors) and the exploitation of several key technological differentiators such as the precise point positioning service (PPP), the potential use of the Galileo signal authentication and the provision of an integrity layer to assess the degree of trust one can associate to the position information provided by the device. Therefore, the three key pillars of the safety-oriented navigation technology provided by ESCAPE engine are: The smart exploitation of different localization data sources to provide a highly accurate navigation solution, including GPS and Galileo dual-frequency measurements, intelligent cameras providing lateral distance to road lane-marks, inertial measurement units, vehicle odometry, PPP corrections and high definition maps; The unique provision of real-time integrity protection levels associated to the location estimates, which express the "degree of usability" of the positioning information for safety-critical applications. The PLs associated to the positions computed relative to the map are fundamental for autonomous driving applications; The full integration of the ESCAPE engine into a vehicle with autonomous driving capabilities, and its test on several different reference paths and environmental conditions. [GRAPHICS] The innovative ESCAPE GNSS Engine (EGE) is close-to-market, its components are organized in a modular architecture and has safety at its core as its specification and design are based on the ISO 26262 recommendations ([2]). ESCAPE was already introduced in [3]. The following sections provide a brief description of the ESCAPE objectives, its design (system, HW and SW) and the positioning and integrity algorithms (see [3] for further details) before describing the tests that have been carried out and the performance results obtained in open-sky, suburban and urban environments.
