The VAMLAV project is a Vinnova FFI-funded project with the following partners: AstaZero, Zenseact, RISE and AI Sweden. The project set out to create a dataset that includes the computer vision sensors that many Advanced Driver-Assistance System (ADAS) and Automated Driving Systems (ADS) vehicles use and complement them with a high-definition (HD)-map over a known geographic area. The VAMLAV dataset includes sensors such as camera, Light Detection and Ranging (LiDAR), Inertial Measurement Units (IMUs), and Global navigation satellite system (GNSS) sensors. This dataset, publicly available at AI Sweden, offers a corresponding HD-map in OpenDRIVE format covering the Rural Road at AstaZero. The dataset includes adverse weather, multiple maps and drives around the track with emulated traffic work scenarios that can occur. Beyond creating the dataset, the project aimed to validate HD-maps by comparing them to other measurement technologies. It delved deeper into localization for ADS vehicles by comparing various measurement campaigns and designing high-accuracy anchor points. This data was later used to validate and update the HD-map. By comparing different measurement systems and samples on the map, the project hopes to increase the trust in the HD-map over a longer time. This data also makes it possible to experiment more within the field of crowdsourced HD-maps from different systems while having an easier time measuring the accuracy of such maps. Another big part of the project was related to safety therefore some data was collected where the project emulates traffic work at AstaZero. This use case is otherwise difficult to test and evaluate due to the stochastic nature of traffic work in real life. Where the system detected the traffic work with the help of map and sensor data and then distributed the information to other cars in the area.

Future automated vehicles and advanced driver assistance systems are highly dependent on sensors to detect their environment as well as robust, accurate, and cost-effective sensor systems for positioning. 

Global Navigation Satellite systems (GNSS) provide a key technology that enables an absolute position estimate and Network-RTK (Real Time Kinematic) has the potential to meet the requirements of cost, accuracy, and availability. This technology is based on correction data being received from a fixed reference station via e.g. mobile communication. Current implementations have been driven by requirements from applications which operate within a limited region for lengthy periods of time, such as surveying and precision agriculture. These applications can tolerate relatively long initialization times and can afford expensive equipment.

The mass market wants to benefit from infrastructure in place for these applications, but the requirements are somewhat different. Problems occur when the device moves from the coverage area of one reference station to another and reinitialization must be made. Consumer devices must also deliver similar performance with inexpensive components. In addition to this, the existing public-sector system for distribution of correction data, in Sweden governed by Lantmäteriet/ SWEPOS, is not designed for handling a large number of clients and efficiently distributing correction data to these clients based on their location.The telecom industry in 3GPP (Third generation partnership project) is currently addressing the need for a scalable provisioning of network RTK corrections. Based on the 3GPP specification, the project aimed to develop, implement, test and demonstrate an efficient distribution system for Network-RTK correction data in order to enable cm-level accuracy GNSS positioning for a large number of mobile platforms e.g. automated vehicles.

The NPAD project has:

• Leveraged the existing Lantmäteriet/SWEPOS GNSS reference infrastructure to implement a virtual network of reference stations that provided coverage over selected test areas suitable for supporting a large number of simultaneous users.
• Implemented a scalable GNSS correction data provisioning based on the ongoing work in 3GPP that provides correction data from the reference network to mobile devices;
• Developed test cases for automated vehicle platforms related to positioning and implemented demonstrators;
• Investigated tools and methods for validating the accuracy of integrated GNSS positioning and navigation systems.

The project was coordinated by RISE Research Institutes of Sweden and involved besides Lantmäteriet and AstaZero the following industrial partners: AB Volvo, Caliterra, Einride, Ericsson, Scania, and Waysure.