A Novel Graph Optimization Approach to Localization (GOAL) for Unmanned Systems


Electronics - Sensors & Instrumentation
Infocomm - Robotics & Automation
Show more >


GOAL is concerned with 2-D plane and 3-D space localization technology for robots. The methodology leverages on ultra-wideband (UWB) aided localization system and can be extended to application with other sensors such as inertial measurement unit (IMU), Wi-Fi, Optical Flow and Cameras.

Unlike existing Filter-based localization algorithms which have unacceptable estimation error in altitude, this technology provides high accuracy in the estimation in altitude without adding other sensors for altitude estimation, which saves cost of localization system. Also, this technology has the function for rejecting Non-line-of-sight (NLOS) measurements, which enables this technology to NLOS environment. The design of flexible moving window in this technology decreases the computational resources and increases the robustness to short duration lost in signal, which enables this technology to be applied to robots with ultra-low power processor.


For the proper functionality and performance of the GOAL algorithm, the following operational setup conditions should be met:

  • UWB distance measurements are required.
  • The working area should be an open space no larger than 16m x 16m x 16m without blockage, and the number of anchors should be no less than 4.
  • The distance measurements must be continuous, general distance measurement error should be less than 20cm. There can be outliers, but outliers cannot be continuous (The time period of continuous outliers should be less than 0.5 second).
  • The setup of anchors should not be coplanar or near-coplanar (the anchors positions should not be on a plane), and the height difference of anchors should be more than 1 meter.
  • For rejecting NLOS measurements, accurate initial position (e.g. the accuracy is within 1/10 x 1/10 x working area) of robot is required. 
  • For example, the working area is 6m x 6m, the error of initial position should be within 0.6m x 0.6m.
  • The update rate of localization (e.g. more than 40 Hz for using UWB module from Time Domain Company) should be set higher than that of UWB.

According to previous experiments using UWB module from Time Domain Company, the accuracy in X, Y and Z is estimated at within 20cm. 


GOAL algorithm can run on an ultra-low power processor carried by Unmanned Aerial Vehicle (UAV). It can also be applied easily for other kinds of robots such as Unmanned Ground Vehicles (UGV), autonomous cars and other robotic applications.

Market Trends & Opportunities

Localization is the most basic capabilities of autonomous robot and plays a significant role in robot-related applications.  This invention provides accurate localization in 2-D plane and 3-D space which is applicable to lower power computational devices. According to market report "Small Drones Market by Application (Defence, Commercial), Size (Micro, Mini, Nano), Payload (Sensor, Telemetry System, Camera, Videoscreen, NBC Detection, SAR, Others), Type, Propulsion Systems, & Region - Global Forecasts to 2020", published by marketsandmarkets, the small drones market is estimated to reach USD 10.04 Billion by 2020, at a CAGR of 11.6% from 2015 to 2020. The commercial segment is projected to grow at a comparatively higher CAGR during the forecast period, primarily owing to huge investments made by major players in the development of small drones for applications in agriculture, transportation, medical, and so on.


  • GOAL algorithm leverages on ultra-wideband (UWB) aided localization system.
  • Provides high accuracy in the estimation in altitude without adding other sensors for altitude estimation.
  • Applicable to robots with ultra-low power processor.
Tactile Sensing Glove
IoT Water Flow Monitoring System
Smart Manufacturing Video Analytics via CCTV
Noninvasive Intracranial Pressure (ICP) Monitor
Worksite Safety Integrated Solution
Integrated Sensor Gateway for Digitisation ( ISGD)
Remote Data Acquisition for Failure Analysis
AIoT-based Real-time Monitoring and Behavioural Pattern Recognition for Human Activities
Secured IoT Gateways
High-Performance Bonding Of Dissimilar Materials Using Nanostructures