Multi domain modeling aids in the movement of automated guide vehicles
Design and control of AGVs, robots, is easier with enhanced modeling software.
With products ranging from a portable and cost-effective system for automating the collection of environmental water data (Kingfisher) to a rugged and easy-to-use unmanned ground vehicle for rapid prototyping (Husky), Clearpath Robotics is progressing in the complex and growing field of robotics. The roboticists at Clearpath specialize in automated guided vehicle (AGV) design and control, and their AVG systems are used by some of North America's leading technical institutions for cutting-edge industrial and military applications. One significant challenge the company faces, given the multifaceted nature of its products, is in the design and modeling of the systems.
“Each interrelated system has to be modeled separately and integrated into the final design,” said Ryan Gariepy, CTO of Clearpath Robotics. “We work with different domains at the same time, and there is a great deal of modeling and prototyping involved in each of the systems.” Software is available for “highly efficient multidomain systems modeling, and it has the potential to completely change the way we design robots."
Physical modeling and simulation software uses a sophisticated symbolic computation engine to handle all complex mathematics involved in the development of engineering models. An intuitive multidomain modeling and simulation environment includes mechanical, electrical, hydraulic, thermal, and signal-flow systems.
From educational robots to all-terrain vehicles comprising drive systems that encompass energy stores and kinematic configurations and actuators that use power transmissions, from payloads to multiterrain systems, designing each robot gives the Clearpath team its own challenges. Moreover, in an effort to maximize efficiency, engineers are constantly trying to consolidate models of all the systems that they work with.
Add to this new industry trends. Previous generations of robots operated in ideal indoor conditions, on concrete or tile floors, whereas today’s all-terrain robots operate in sand, mud, and water. Therefore, there is an increasing need to include capabilities such as sensing, processing, and communication in modern robots. Plus, there are additional mobility challenges like balancing the ability of the robot to overcome obstacles with quick movements on a flat surface. All these challenges mean new hurdles in design, modeling, and, of course, performance. They require a lot of complicated models with several design iterations, hindering engineers’ ability to explore further avenues in cost analysis, modeling, and designing.
3-D multibody dynamics and multi-domain modeling capabilities help overcome these challenges. Other types of software “still need to be manually coupled into our models and design process. Using those tools also means that we have to have experts in each domain who will take care of that particular component.” The software eliminates the hurdles. “Everybody is able to work on a single model. This is why we're very excited,” said Gariepy.
Clearpath frequently needs to design robots based on a new set of requirements, different surface conditions, robot size and shape, or performance criteria. The software models can be highly parameterized, so they provide an ideal rapid prototyping environment for creating new designs. They build base models in which many elements that would be traditionally assigned a fixed value are actually implemented as symbolic parameters, so that they can easily explore the effects of changing these design elements without the need to create a new model. Using these base models, they can adjust the parameters to meet the new conditions, and immediately get a starting point for the new design, which they can then further refine.
It’s useful to work with fewer equations than others, using more of a symbolic approach, which “takes care of the grunt work and computes quickly. It's all that we're looking for in a tool,” concluded Gariepy.
- Paul Goossens is vice president, applications engineering, Maplesoft. Edited by Mark T. Hoske, content manager CFE Media, Control Engineering, Plant Engineering, and Consulting-Specifying Engineer. Hoske can be reached at firstname.lastname@example.org