Sawppy is a motor model, its layout and proportion imitates the design idea of Mars Rover, which is faithfully reproduces the real roamor suspension kinematics, aimed to be the hardware platform of future self-run software projects. Continue to explore the backyard, Sopu!
Sawppy The Rover is inspired by the JPL Open Source Rover project. Most of the differences between Sawppy and its JPL inspiration is to reduce cost and reduce complexity. JPL's roaming car is designed for education, and is assembled by a school team to lay a solid foundation for structured courses. Sawpy is more suitable for personal enthusiasts like me, they are willing to fix and willing to make some compromise to reduce costs.
The budget is $ 500, and it is necessary to reach the following changes:
Mobile: Instead of using a gear motor with an encoder managed by the Roboclaw motor controller, but use a serial bus servo motor to complete the movement of the wheel.
Construction method: instead of using an Actobotics construction system, Sawppy will be manufactured by Misumi 15mm 3 series aluminum profiles and is connected by 3D plastic parts.
These two main design goals can be summarized as: small ERVO a CTUATED W ¯ high heels, P rINTED I NTERCONNECT ˚F or xtrusion. The first alphabetic word Sawpife caused the nickname "Sawppy".
When I created Sawppy, my first goal is to provide interesting things for robotics. Target crowd is a high school student and above, which means creating a separate software stack and is concentrated to learn and modify.
In order to meet the ambition of Sawppy developers, one of the future works is to write the necessary modules to drive SAWPPY through the open source robot operating system. (ROS) This is a more powerful platform that accesses the module created by robot researchers, but it is not easy for robot beginners. Since then, I will continue to play with ROS, and I have never reached the proficiency I needed.
Therefore, I was very excited about Rhys mainwaring's Curio wanderer. Curio is SAWPPY brothers and sisters, the body is basically the same, but it is running a completely different software stack based on ROS Melodic. When browsing the Curio code repository, I see is not just a set of nodes running physical roaming vehicles, but it also includes two important contributions to intelligent roaming vehicles.
Simulation
Curio Rover Today, the intelligent robot has a common problem: the growing machine learning algorithm requires multiple iterations, and it takes a long time to run them on physical robots. More importantly, in real life, Sawppy and their brothers and sisters are very slow. RHYS has converted SAWPPY CAD data and converts the physical form of the Gazebo robot simulator used by ROS researchers and all joint motion. You can now work on the intelligent roaming car in the virtual world before adapting to the real world.
Mobile station mileage table
One of the challenges I realized (but I don't know how to solve) is to calculate the mobutometer. The LX-16A servers used on SawPPY can return to the wheel position, but can only be included in the entire 360 degree arc of the arc of approximately 240 degrees. Being out of this range, location data will become noisy and unreliable.
RHYS has been tryed to overcome this problem through the encoder filter, which learns when the identification of the servo position data is not reliable. This constitutes the basis of the system of calculating the mileage, which can be used well with existing hardware, and use additional Arduino can even be faster.
ROS software stack for SAWPPY
Some people have to ask me for Sawppy ROS software, I am very happy that RHYS is brave in the challenge and gives back this work to the community. I encourage all the hopes of the Sawppy builders who want ROS carefully study RHYS, and do their best to make contributions. As an initiator of ROS, I will work side by side and learn from this project and try to run it on my mobile station.
Be
Article Source HackADAY
Our other product:
