Seoul Robotics (CEO Hanbin Lee) announced on the 11th that it has acquired a patent for a ‘virtual space-based vehicle control system using sensor data’ from the European Patent Office (EPO).
This patent concerns a 3D object recognition AI technology based on deep learning. The core of the technology is a multimodal AI technology that analyzes 3D spatial information collected from sensors in real time 10 to 20 times per second through a neural network-based model and extracts spatiotemporal features to detect and track objects.
Seoul Robotics explained that in order to solve the problem that existing rule-based autonomous driving systems only work in limited environments, they have learned various edge case data from various industries over the past eight years through deep learning to implement highly accurate object recognition even under harsh weather conditions.
In addition, the company's independently developed deep learning system compresses 3D point clouds to achieve fast inference speed and high accuracy, and supports mass data learning through a ResNeXt-based backbone network and automatic labeling system. This technology replicates real-time operating environments such as ports, factories, and logistics complexes to efficiently integrate and control multiple vehicles, and enables unmanned autonomous driving even in bad weather such as heavy snow or heavy rain.
Lee Han-bin, CEO of Seoul Robotics, said, “This acquisition of a European patent is the result of international recognition of the originality and advancement of our technology,” adding, “Europe, in particular, is a market with a high concentration of automobile manufacturers and logistics automation powerhouses, and this will serve as an important foundation for local advancement and expanded cooperation.”
He added, “Securing exclusive rights to unmanned transport solutions and fleet management systems in Europe will also strengthen our negotiating power for global partnerships.”
- See more related articles
You must be logged in to post a comment.