Page 18 - Gi flipbook_May 2019

machine learning

for different ground types; including
difficult to excavate, hard-packed clay.
Using computer simulation, finite
element analysis (FEA) and rapid
prototyping, coupled with extensive
laboratory testing, multiple supersonic
air nozzle designs were developed,
which substantially improved upon
the existing air lance nozzle designs,
enabling greater efficiency.
In the RRES’s design, air provided by
a standard utility compressor is injected
into the divergent section of the nozzle
at subsonic speeds (below Mach 1). As
the air moves from the divergent section
to the throat, it undergoes compression LAB TESTING OF EXCAVATION
(due to the area change), causing the
velocity to increase and the pressure to
decrease. When the volumetric flow rate
and the pressure reach a critical point
called ‘choked flow’, air at the nozzle’s
throat will surpass sonic speed (Mach 1),
thus creating supersonic flow.
The RRES’s supersonic air nozzles
were designed for integration into a
compact, custom-designed vacuum
excavation head, which enables
effective agitation and removal of soil
without the risk of damage to buried
assets. When mounted to the robotic
arm, this excavator head can be
controlled manually or with varying
degrees of autonomy. The figure shows THE RRES HAS A STEROO VISION SYSTEM
the compact excavator mounted at
the end of the robot arm in both the location and size of exposed utilities costs, setting a new standard for the
simulation environment as well as the during the excavation, must be known future of roadworks. Additionally, the
lab environment. during the excavation process. Using RRES is being developed as a platform
this technique, a 3D point cloud of the for future ideas and the RRES team has
Post-excavation sensing entire excavation can be generated plans to develop it further to support a
When human operators excavate, they in near real-time. With the proper variety of cross utility and construction
use their eyes and senses to control presentation of this dataset to the applications.
backhoes and other large construction operator and to the computer control
equipment. In order to embed the RRES systems, it can be used as a tool for ULC Robotics is a leader in product
with the ability to ‘see’ its environment, manual or automatic control of the development and engineering services for
the RRES team is developing 3D excavator head, sensors and road the natural gas, energy, industrial and
visualisation techniques to capture cutting tooling during the operation. electric distribution markets. For more
3D point clouds of the excavation and information, visit: www.ulcrobotics.com
surrounding site. What’s next?
A point cloud is a set of data points Over the next two years, the RRES
that represent points in 3D space. project is focused on developing and
These can be used for measurement, deploying the next generation
navigation and to generate extremely of excavation technology
accurate 3D models of environments. with a goal of reducing
In the RRES system, point clouds are the risk associated with
being paired with proprietary software to labour-intensive operations,
measure the depth of the excavated hole public and environmental
and to support the control/guidance of disruption, and the risk
the arm and excavation tooling to enable of damaging unmarked
safer operation, more accurate path third-party infrastructure.
planning and greater efficiency. By eliminating the need for
To support this goal, a stereo vision traditional equipment and
system has been developed to capture processes, the system promises
high-resolution 3D point clouds a range of benefits such as safer
from the excavation environment. working conditions, decreased
Information such as the dimensions traffic disruption and substantial
and texture of the excavated soil, reductions in carbon emissions and
CHAINSAW TOOL FOR CUTTING THE ROAD SURFACE

18

ULC Robotics.indd 3 16/04/2019 19:41