summary
Science fiction has inspired many dreams of becoming engineers for countless young people, perhaps building a “bot” of a day, especially in films featuring robots such as R2-D2 and Wall-E. Sci-fi inspired engineers continue to do great things, but few people actually build their own personal animatronic robots. In particular, there is no match for the more famous on-screen robo counterparts.
Growing up, I consumed science fiction and enjoyed frequent family trips to theme parks. However, unlike the general public, certainly like many readers of this publication, I tended to find the animatronics and control systems behind these attractions more appealing than vehicles. Disney’s carousel advances were particularly interesting as its robotics seemed far beyond the era.
Later I received technical training in Air Force and University Credits through Augusta State University and other institutions, and also spent many years of practical work in the industrial automation profession. But I always wanted to develop my own personal robot. Building a full-size animatronic robot, even with the right inspiration, skills and determination, is a large, non-estimated venture. This is how I have done this dream over the years.
Automation Inspiration
My ambition was to create the Hubert-full name Hubert Evolution 2, or HEV2, to be reminiscent of the charismatic Johnny 5 from the 1986 film Short Circuit (see Figure 1 at the top of the article). This is not a formal educational project, but rather a much more way to get out of the book instead, but as the build evolved, I was able to receive school credit for certain elements of the build. This work began in 2006 when I was a student, and continued as a part-time project until around 2020, and continued development for about 15 years.
When I made professional advancements in the OEM automation industry, I gained skills in the workplace and then acquired skills in the Hubert project, which translates back and forth. I love robotics and automation, but I would never argue that this project is easy and countless hours spent on design, manufacturing and programming. However, the learning opportunities were incredibly valuable. My strong mechanical aptitude and various technical experience have been helpful and this project includes a lot of research and testing.
The design started small, but I knew it would be expensive to expand to full size. I started with a kind of cart or tank-like frame made from aluminum extrusions as the basis for building all the automation. I knew I wanted Hubert to have a mobile torso, limbs, head and other features. Sometimes parts are available cheaply from eBay, but this hasn’t always been the case. Consumer grade hobby servos were also useful.
When investigating automation techniques, we quickly concluded that we should avoid extensive experimentation in custom board-level electronics and coding, and to use commercial ready-made (COTS) products to the maximum possible extent. For these and other reasons, I quickly graduated from using more competent control electronics, software, electromechanical hardware, and pneumatic pressure.
Industrial-GRADE Tech supports animatronic features
My research has led me to rely on AutomationDirect to evaluate and obtain various automation and control components. Some of the earliest Hubert components were clearly consumer grade discoveries, but the movement incorporating industrial grade automation direct hardware and software was a natural fit. The website has a wealth of technical information, support videos and other resources that helped you select components to implement new programming.
Hubert’s design was a process of continuous improvement. For example, I started with a small hobby servos that requires a DC servo controller. I discovered that to command this controller, I could send serial strings and generate a wide range of motion. At one point I landed at Stepper Motors and drove as another economical way to create movement.
Initially, we did not use a dedicated motion controller. Instead, we found that the AutomationDirect Click Koyo Programmable Logic Controller (PLC) is easy to use, offers excellent flexibility at the price, has free software, generates serial strings, and is ideal for interacting with drives using fast input/output (I/O) signals. As I/O counts continued to grow, I improved my organization using pre-riding cables and connectors on AutomationDirect Ziprink.
As the project progressed, I moved using air pressure. Air pressure is less accurate than servos and steppers, but it is a powerful way to achieve movement. Linear pneumatic actuators combined with position feedback and nitra valves provide the best way to move Hubert’s limbs, and the compressibility of the air allows them to move with a more organic feel than electrical options.
Soon I was trying out a much more advanced algorithm. I was already using a considerable amount of geometry and trigonometry in my code. We began using proportional integration (PI) controls of PLC origin and experimented with timed loop controls. The goal was to produce smooth, natural movement with the ability to hold the pneumatic cylinder in its ordered position within a few millimeters. However, as codebase and device counts increased, Hubert required more brain power, so he combined AutomationDirect’s 3000 Modular PLC with the AutomationDirect C-More Human-Machine Interface (HMI) to provide user visualization and data/command input methods (Figure 2).
I have found it very useful to add more actuators and drive them with a full proportional integration (PID) control, especially after adding different gains. However, when this stopped meeting my needs, I started writing code that implemented the maths needed for fuzzy logic.
At one point, Productivity3000 alone contained almost a megabyte of ladder logic! Additionally, the final development included an AutomationDirect Click PLC, eight Arduinos, and two Raspberry PI microcontrollers. Part of the learning process is gaining experience with open source controllers, including the industrialized Arduino compatible controller AutomationDirect ProductivityOpen.
Productivity3000 is the main supervisory controller, interacting with other controllers that use C++ and Python to perform additional detailed tasks, exchanging data and interacting with other controllers. The OpenGL and OPENCV routines enable Hubert’s light feature, including 1.5 inch OLED animation eyes, while the voice recognition and output hardware allow interaction with bystanders.
Today, Hubert’s Side’s industrial style control panels have PLC and HMI, with joysticks, illuminated push buttons and emergency stop (E-stop) buttons. Most of the other electrical and automated items are sourced from AutomationDirect. Hubert is about 7 feet tall and weighs well over 1,000 pounds (Figure 3).
Build a bot in your career
Hubert’s development was a labor of love for 15 years. Built primarily with industrial automation components, it was an excellent learning experience with crossovers that are very relevant to my professional work. Has Hubert finished? Not much; future plans include updating HMI, adding Ethernet/IP control valve banks, integration of a more advanced productivity motion PS-AMR4 4-axis motion controller, manufacturing additional components and external shells using a 3D printer and CO2 laser.
Recently I bought some potentiometer-equipped gloves and integrated them so I could use my hand movements to teach and record the movements Hubert would perform later. We’ve also worked to create a “scene” of small movements that can be called if necessary.
The immediate accessibility and ease of use of AutomationDirect products is the basis of this project and makes it a great platform for anyone who likes to innovate or is aiming to innovate using automated technology. Hubert hopes that he is urging Hubert to create his own new projects to open up new careers and other opportunities.
All numbers courtesy of Justin Flynn
Have you enjoyed this amazing article?
To read free articles, check out our free e-newsletter.
Subscribe
