QRTECH is continously engaged in research projects focusing on cutting edge technologies that will shape the future of automotive industry. Competence development through research projects helps us maintain our position as leading-edge experts in the development of embedded systems for automotive.
The development of safe autonomous functions that live up to industry safety standard are, despite significant progress made by innovative companies, considered to be in an early phase. Currently autonomous and driver assistance functions rely on sensor input from RADAR, LIDAR, Ultra-sonic, IR, and camera sensors. Algorithms that take safety critical decisions based on this input data are still under development and the automotive industry is starting to acknowledge the potential for using deep learning techniques.
Autonomous driving using control algorithms and neural network models is today being developed using bulky PCs filling the trunk of concept cars.
The problem with the current approach is that the hardware and low-level software in the PCs are far from ready for use in the safety critical automotive domain. In fact, an automotive grade computation and communication platforms with sufficient real time performance does not yet exist.
Using our QRx automotive grade platform we are adressing the major upcoming challenge for autonomous driving for series produced cars; finding electrical architectures that have the computational resources and high speed communication interfaces capable of supporting performance demanding control algorithms and neural networks in real-time.
DEEP LEARNING AND NEURAL NETWORKS FOR OBJECT RECOGNITION
Deep learning has been proven extremely effective across many application domains. It is transforming the way computers achieve perceptual tasks, such as pattern recognition and computer vision. In the context of autonomous cars, pattern and object recognition are expected to be unavoidably crucial functions in order to achieve safe behavior while interacting in complex environments with human drivers and pedestrians. In essence: the complexity of the driving situation makes it is extremely difficult to cover the problem space using the classical deterministic approaches to algorithm generation.
QRTECHs research aims to take deep learning technology and control algorithms for autonomous driving a step closer to the market. By experimenting with deployment of neural network and control algorithms onto an open scalable automotive grade prototyping platform, important insights on the enabling technologies for performance and dependability in future series produced cars will be gained. The project will consider the architectural distribution of computational resources and communication interfaces as well as run-time data and services needed to deploy neural networks on automotive grade electronics.
Safe inductive charging, balanced between performance and personal safety. Small magnetic fields to ensure the personal safety of users, while maintaining a high system efficiency.
Qrtechs inductive charging system features
- 3.5 kW(230 V, 16A, Single Phase)
- 80-90 kHz operating frequency
- Air gap 250 mm
- Transmitter coil geometry 500x500x30 mm
- Receiver coil geometry 400x400x30 mm
- Active shielding coil control
- Foreign object detection
- Bluetooth communication
- User control via Android app