Non-iterative Simultaneous Localization and Mapping (NI-SLAM) is a patented method for position tracking and environment dense reconstruction, using affordable sensors such as depth camera and inertial sensors. Compared to traditional SLAM methods, the superior real-time processing speed of NI-SLAM is achieved by an innovative method of data registration that performs point-cloud matching in Fourier Domain.
NI-SLAM relies on a high-speed inertial sensor, coupled with dense point-cloud data, which is acquired either by passive RGB-D camera or active depth scanner. In order to perform point cloud matching while avoiding costly iterative routines of traditional SLAM methods, our innovative algorithm leverages on Kernel Cross Correlation, which makes clever use of certain special characteristics of circulant matrices for fast machine learning.
Furthermore, NI-SLAM has been experimentally compared with other traditional SLAM methods, demonstrating superior performance on both high and low computation platforms.
Adoption of robots is becoming a new norm in virtually all aspects of the society. Not only in labour-intensive industries such as Manufacturing and Logistics, traditionally “human-dominant” sectors such as Nursing, Food & Beverages and Health Care also see increasing use of robots to compensate for the ageing population and manpower shortages.
Interestingly, demands from emerging markets have gradually shaped new designs and functions of modern robots. New generations of robots, made to assist and participate in daily human activities, has to be more mobile, more user-friendly, more interactive and, most importantly, smaller in size. The physical constraints induce various hardware limitations, especially on computers and batteries. Therefore, there are rising market opportunities for technologies that could help robots leverage well on the limited amount of onboard hardware resources. NI-SLAM is one of such potential technologies.
Advantages of NI-SLAM, including high processing speed and low hardware requirements, allow the user to deploy SLAM technologies on small platforms such as drones and mobile robots, which are increasingly affordable and popular in various industries such as Logistics, Food & Beverages, Health Care, and Transportation. Moreover, because of smaller power consumption and less demanding hardware requirements, customers could further reduce both their acquisition cost, in large-scale deployment and maintenance cost in the long run.