Automatic Map Calibration for Markerless Localization (AMaCal)

Abstract:

The goal of the proposed project is proof of concept evaluation and market potential analysis of an innovative
method for improving autonomous forklift navigation in warehouses. If proven feasible, the proposed method
would significantly reduce the commissioning costs and increase the flexibility of operations in warehouses
equipped with autonomous forklifts. This would remove a significant barrier on the way towards wider adoption of
fully autonomous vehicles in logistic operations.
Autonomous forklifts are equipped with safety laser distance sensors, which bring the vehicle to an immediate
stop in case an obstacle or person is detected in its path. The proposed solution uses distance measurements
from the safety lasers for building a map of the environment. The map is built fully autonomously, without human
operator involvement. The generated map is then automatically calibrated against a CAD floor plan of the
warehouse. CAD floor plans are readily available, as they are commonly used for planning warehousing
operations. The calibration procedure is performed by determining a set of coordinate transforms matching the
corresponding segments between the two maps using The Iterative Closest Point (ICP) method. If proven
feasible, this automated calibration procedure would enable fully autonomous warehouse commissioning, which
would result in significant savings.
The main objectives of the project are the implementation of the automated calibration procedure and its
validation. In addition to validating the technical feasibility of the proposed solution, it is our objective to evaluate
its market potential and perform preliminary steps towards IP protection.
The implementation of the project is planned in three phases. In the first phase, we will implement the prototype,
and its functionality will be evaluated in laboratory conditions. In the second phase, the functionality will be
evaluated in industrial conditions, with support of partner companies. The third phase will be dedicated to market
analysis, preliminary IP protection and the development of a commercialization strategy.
Proof of concept evaluation is necessary in order to determine the actual technical feasibility of automated map
calibration, thereby lowering the risk of further development of a novel autonomous forklift navigation product.
The expected results are a fully functional software prototype and a strategy for further development and
commercialization

Project duration:
July 2016 – June 2017

Principal investigator:
Dr.sc. Damjan Miklić