ROBOTIC ROADWORKS & EXCAVATION SYSTEM

Transcription

1 ROBOTIC ROADWORKS & EXCAVATION SYSTEM Ali Asmari PhD Senior Robotics Engineer Artificial Intelligence & Machine Learning

2 Challenges With Traditional Excavation Methods Dangerous Underground Assets Road Closures & Traffic Third Party Damages Equipment Failures Storms & Inclement Weather Public & Business Disruption Third Party Repair Work Excessive Landfill & Debris

3 Robotic Roadworks & Excavation System The Method Through two phases of feasibility studies ULC and SGN developed a conceptual method for using robotics to: Automate Improve Safety Decrease the Costs of excavation operations around buried distribution and transmission piping

4 Robotic Roadworks & Excavation System Scope of Work Element 1 - Development Of Major Subsystems Robotic System Below-ground Sensing Excavation Tooling Element 2 - Integration Of Element 1 Subsystems And Initial Testing Integration & Field Testing 1 Element 3 - Mobile Operations, Fittings, Tooling And Support Vehicle Mobile Operations Tool Changing Fitting Installation Support Vehicle Element 4 - Complete Integration And Live Field Test Final Integration & Field Test 2 Fitting Installation Tooling Support Equipment Tool Changing System Robotic Arm RRES Excavation Tooling Mobile Platform Belowground Sensing System Computing System

5 Robotic Roadworks & Excavation System Operation Procedure Excavation Underground Sensing Core Drilling Core Removing Vacuum Excavation Operation Future Development of Autonomous Operation on the Exposed Pipes Universal Access Fitting Reinstatement Backfilling Soil Tamping Soil Core Reinstating

6 Excavation Soft Touch Breaking the soil and excavation in one operation Avoiding any potential damage to the buried utilities Enabling the application of core-and-vac excavation to previously prohibited sites Reducing the time it takes to carry out the operation

7 RRES Robotic Arm Operations Cutting The Road Surface In order to efficiently cut the road surface ULC is looking at alternative techniques such as: Milling out the concrete core Using a concrete chainsaw to cut the core Cutting the core using an end mill To accurately study and measure the requirements for cutting the concrete road surface using these methods, ULC is building a 3 axis CNC machine with appropriate force and torque feedback, capable of articulating different tools of different size and shapes.

8 Robotic Roadworks and Excavation System Below-Ground Sensing Pre-excavation surface scanning RRES will scan the surface of the marked out location to detect underground pipelines, cables and obstacles in the first layer prior to breaking ground Layer-by-layer scanning during excavation Sensors used after each layer of soil is removed RRES system will alert operator if an unknown object is identified Machine vision recognizes assets and avoids contact Machine vision cameras and sensors will allow the system to recognize pipelines, cables and objects as layers of soil are removed Support the installation of the Universal Access Fitting (UAF)

9 Goal: Below Ground Sensing Pre-Excavation Evaluation of (1) Capabilities and (2) Limitations of each sensor technology in detecting various buried utility lines of different material Soil Box Specifications: Two compartments filled with Sand and Clay to represent two extreme cases of soil type Extreme cases of each utility line was selected to be buried in each compartment Pipe type Typical diameter Minimum diameter Gas pipe (plastic) 75mm, 90mm 20mm Gas pipe (metal) 4 inch 0.75 inch Electrical line 2 inch, no conduit 0.5 inch, 10mm Telecomm cable/fiber 4 inch, w/ conduit 10mm optic Cable TV (coax) 4 inch, w/ conduit 10mm Water pipe (plastic) 90mm, 75mm 25mm or 15mm Less common: 110mm, 125mm Water pipe (metal) 4 inch 0.5 inch

10 Below Ground Sensing Pre-Excavation Sample of data collection and presentation from the soil box, exposing different utilities of various sizes

11 Below Ground Sensing Stereo Vision ULC has developed a stereo vision system for extracting depth and pattern information Using pattern matching the overlapping section in left and right images are detected By fusion of the data from left and right images, depth information can be extracted from each datapoint Left Right

12 Below Ground Sensing Stereo Vision Images are captured from two synchronized cameras Using image processing techniques, depth information can be extracted from the target

13 Thank You For Attending! For Questions & Further Information Please Contact Ali Asmari PhD Program Manager Robotic Roadworks & Excavation System

14