Global Earth Monitor project (GEM) is addressing the challenge of continuous monitoring of large areas in a sustainable cost effective way. The goal of the project is to establish a new disruptive Earth Observation (EO) DATA-EXPLOITATION MODEL which will dramatically enhance the exploitation of Copernicus data. For the first time a continuous monitoring of the planet on the global/regional scale will be enabled for a sustainable price. Disruptive innovations are planned in the technology and in the methodology domain, where a proprietary concept of Adjustable Data Cubes (a combination of static and dynamic data cubes) will be developed and integrated with EO-oriented open source Machine Learning (ML) framework EO-LEARN. During the project EO-LEARN will be upgraded to consume ML technologies from widely accepted ML frameworks and to adapt/evolve them to specifics of EO-data interpretation. Modern ML technologies and approaches (GAN, RNN, LSTM, Attention & Bayesian Deep Learning, Curriculum Learning, Incremental learning, Meta-learning, Hybrid modelling) will be combined to construct GLOBAL, SCALE-INDEPENDENT interpretation models with the special focus on CAUSALITY and CHANGE DETECTION. Technological and Methodological innovations will be combined into a unique CONTINUOUS MONITORING PROCESS. The process, based on seamless combination of data interpreted with sub-resolution, native resolution and super-resolution methods, will deliver optimal combination of Processing/Storage costs – enabling continuous monitoring of large areas for just a FRACTION OF CURRENT COSTS. The concept of continuous monitoring will be validated through the development of five specific use-cases and through their employment in a 10-month demonstration - operational continuous monitoring of 10 MIO square km area.

Lane-level positioning and map matching are some of the biggest challenges for navigation systems. Although vehicle telematics provide services with positioning requirements fulfilled by low-cost GNSS receivers, more complex road and driver assistance applications are increasingly been deployed, due to the growing demand. These include lane-level information as well as lane-level navigation and prioritised alerts depending on the scenario composition (traffic sign, navigation instructions, ADAS instructions). These applications need a more accurate and reliable positioning subsystem. A good example of these new requirements can be witnessed in the increasing interest in navigation at lane-level, with applications such as enhanced driver awareness, intelligent speed alert and simple lane allocation. As well as the accuracy of positioning data being a big driver, there is also a question around the adaptability of navigation systems to these applications. This depends firstly on the availability of an accurate common reference for positioning (an enhanced map) and secondly, on the level of the provided pose estimation (integrity). However, neither the current road maps nor the traditional integrity parameters seem to be well suited for these purposes. Delivering lane-level information to an in-vehicle navigation system and combining this with the opportunity for vehicles to exchange information between themselves, will give drivers the opportunity to select the optimal road lane, even in dense traffic in urban and extra-urban areas. Every driver will be able to choose the appropriate lane and will to be able to reduce the risks associate with last-moment lane-change manoeuvres. inLane proposes new generation, low-cost, lane-level, precise turn-by-turn navigation applications through the fusion of EGNSS and Computer Vision technology. This will enable a new generation of enhanced mapping information based on crowdsourcing.

Cloud-LSVA will create Big Data Technologies to address the open problem of a lack of software tools, and hardware platforms, to annotate petabyte scale video datasets. The problem is of particular importance to the automotive industry. CMOS Image Sensors for Vehicles are the primary area of innovation for camera manufactures at present. They are the sensor that offers the most functionality for the price in a cost sensitive industry. By 2020 the typical mid-range car will have 10 cameras, be connected, and generate 10TB per day, without considering other sensors. Customer demand is for Advanced Driver Assistance Systems (ADAS) which are a step on the path to Autonomous Vehicles. The European automotive industry is the world leader and dominant in the market for ADAS. The technologies depend upon the analysis of video and other vehicle sensor data. Annotations of road traffic objects, events and scenes are critical for training and testing computer vision techniques that are the heart of modern ADAS and Navigation systems. Thus, building ADAS algorithms using machine learning techniques require annotated data sets. Human annotation is an expensive and error-prone task that has only been tackled on small scale to date. Currently no commercial tool exists that addresses the need for semi-automated annotation or that leverages the elasticity of Cloud computing in order to reduce the cost of the task. Providing this capability will establish a sustainable basis to drive forward automotive Big Data Technologies. Furthermore, the computer is set to become the central hub of a connected car and this provides the opportunity to investigate how these Big Data Technologies can be scaled to perform lightweight analysis on board, with results sent back to a Cloud Crowdsourcing platform, further reducing the complexity of the challenge faced by the Industry. Car manufacturers can then in turn cyclically update the ADAS and Mapping software on the vehicle benefiting the consumer.