Technology Update: Leg-wheeled hybrid mobile robot
Flexible programming helps in developing an energy-efficient leg-wheeled hybrid mobile robot that can drive quickly and smoothly on flat terrain and stably negotiate natural or artificial uneven terrain.
Integrated hardware and software helped with a mobile robotic design that uses the best attributes of legs and wheels, two widely adopted methodologies for ground locomotion.
Legs of most ground animals are agile, powerful, and capable of moving animals smoothly and rapidly on uneven, natural terrains. Humans invented wheels for specialized locomotion on flat ground, providing excellent power efficiency and smooth travel at high speeds on flat ground, exceeding capabilities of legged locomotion.
Combining wheeled mobility on flat ground and legs on rough terrain, the Bio-Inspired Robotic Laboratory (BioRoLa) team at National Taiwan University aimed to design a leg-wheeled hybrid robot to handle flat or rough terrain.
Compared to most hybrid platforms, which have separate mechanisms and actuators for wheels and legs, the Quattroped leg-wheeled hybrid mobile robot uses a “transformation mechanism” that deforms a specific portion of the body to act as a wheel or a leg.
From a geometrical point of view, a wheel usually has a circular rim and a rotational axis located at the center of the rim. The rim contacts the ground, and the rotational axis connects to the robot body at a point hereafter referred to as the “hip joint.” With most wheeled locomotion on flat ground, the wheel rotates continuously and the ground-contact point of the wheel is located directly below the hip joint with a fixed distance. In contrast, in legged locomotion the leg moves in a periodic manner without specific geometrical configuration between the hip joint and the ground-contact point. The relative position of the legs varies frequently and periodically during locomotion.
Shifting the hip joint out of the center of the circular rim and changing the continuous rotation motion to other motion patterns implies the locomotion switches from wheeled mode to legged mode. This provided design motivation for a mechanism that directly controls the relative position of the circular rim with respect to the hip joint so it can generate wheeled and legged motions. Because the circular rim is a two-dimensional object, adding a second degree of freedom (DOF) can adjust the relative position of the hip joint to the center of the circular rim along the radial direction. Two DOF motions are orthogonal to each other. The same set of actuation power can be efficiently used in wheeled and legged modes.
Robot computation is mainly powered by a 400 MHz real-time embedded control system operating at a loop rate of 1 kHz with a gate field-programmable gate array (FPGA) embedded chassis running at 10 kHz. The real-time control system microprocessor communicates to a remote-control (RC) PC laptop via standard IEEE 802.11 wireless protocols. The FPGA connects to digital and analog I/O modules, which connect to robot sensors and actuators.
Robot sensors include temperature sensors on the motors and power amplifier chips for health monitoring; voltage and current measurement sensors for power management; Hall-effect sensors for leg-wheel configuration calibration; a 6-axis inertial measurement unit (IMU) and a 2-axis inclinometer for body state measurement; and three infrared (IR) distance sensors to measure ground clearance. Various sensors, such as GPS, vision, and laser ranger, are also used to improve the robot’s perception ability. Actuators on the robot include eight dc brushed motors for driving the robot, two high-torque remote-controlled servos for front leg-wheel turning, and four small remote controlled servos and four small dc brushed motors for leg-wheel switching.
Software, three cores
Three computation cores (PC, real time, and FPGA) running object-oriented control software are responsible for different tasks. A remote-control PC operated by the user exchanges only essential information with the microprocessor of the real-time controller, such as high-level commands to drive the robot in different modes, passing back crucial motor and electronic status for health monitoring and logging state data.
Most computation is executed within the onboard real-time processor. Some algorithms that require high-speed signal exchanges are compiled within the FPGA, such as proportional-integral-derivative (PID) control for the dc motors, encoder readings, and pulse-width-modulation (PWM)-based RC servo commands. The robot is programmed with various state machines, and each state represents one of the robot’s particular operating behaviors.
After the robot is powered on, motors are calibrated to define the absolute geometric configurations of the two active DOFs on each leg-wheel. Calibration is achieved by matching the relative position between Hall-effect sensors installed on the body and magnets mounted within the leg-wheel. The calibrated robot can be operated in legged mode or in wheeled mode, depending on the current rim configurations (that is, wheel or half-circle leg). Leg-wheel switching transforms the leg-wheel configuration.
In wheeled mode, the robot can stand, drive, and sit. Standing and sitting are two transient states to bridge the initial on-the-ground configuration and the driving behavior. When driving, the forward speed and turning rate are continuously adjustable. In legged-mode operation, standing and sitting behaviors also are included for transient states. After the robot stands, it can perform various behaviors, including walking, trotting, step crossing, bar crossing, and stair climbing.
Hardware, software benefits
Robots, in general, are high-DOF complex systems. Successful robot development requires time and effort to properly integrate various mechanical, electrical, and computer systems. For the BioRoLa team at National Taiwan University (mainly students with mechanical engineering backgrounds), a reliable, modular, easy-to-use, and well-integrated mechatronic system suitable for rapid prototyping helped robot development.
The graphical programming interface makes it easy for students to construct logic flow followed by coding and to understand the programs created by other developers. A compact, durable, and modular controller helps mobile robot development where the size, weight, performance, and learning time are important factors. Software and hardware compatibility reduces the time and efforts of developers in performing system integration.
On the hardware side, various sensors are being integrated into the current mechatronic system to improve the perception capabilities of the robot. On the behavioral side, legged behaviors are being refined with a closed-loop feature to improve robot mobility on various challenging terrains and to explore dynamic legged gaits.
At a glance: Leg-wheeled hybrid mobile robot application, software and hardware
NI LabVIEW software, CompactRIO hardware, and I/O modules are used to rapidly integrate mechanical design, mechatronics, and programming for a functional robot prototype that combines the best attributes of wheels and legs for locomotion. The system includes the following NI hardware and software.
- Computation is mainly powered by a 400 MHz NI cRIO-9014 real-time embedded control system operating at a loop rate of 1 kHz with an NI cRIO-9104 3M gate field-programmable gate array (FPGA) embedded chassis running at 10 kHz.
- FPGA connects to NI 9205 and NI 9264 analog I/O modules and NI 9401 and NI 9403 digital I/O modules, which connect to robot sensors and actuators.
- Three computation cores (PC, real time, and FPGA) run LabVIEW 8.6. A remote-control PC operated by the user exchanges essential information (high-level commands to drive the robot in different modes) with the microprocessor of the real-time CompactRIO controller. It passes back crucial motor and electronic status for health monitoring and logging state data.
- NI LabVIEW graphical programming interface makes it easy for students to construct logic flow followed by coding and to understand the programs created by other developers.
- Compact, durable, and modular CompactRIO system is extremely suitable for mobile robot development where the size, weight, performance, and learning time are important factors.
- Well-defined integration between LabVIEW and NI hardware significantly reduces the time and effort of developers in performing system integration.
The authors thank NI Taiwan for kind support in equipment provision and technical consulting.
- Authors, from Department of Mechanical Engineering at National Taiwan University, are Pei-Chun Lin, Shen-Chiang Chen, Ke Jung Huang, Shuan-Yu Shen, and Cheng-Hsin Li. Edited by Mark T. Hoske, Control Engineering, www.controleng.com
AIMing for Automated Vehicles blog www.controleng.com/blogs
|Search the online Automation Integrator Guide|
Case Study Database
Get more exposure for your case study by uploading it to the Control Engineering case study database, where end-users can identify relevant solutions and explore what the experts are doing to effectively implement a variety of technology and productivity related projects.
These case studies provide examples of how knowledgeable solution providers have used technology, processes and people to create effective and successful implementations in real-world situations. Case studies can be completed by filling out a simple online form where you can outline the project title, abstract, and full story in 1500 words or less; upload photos, videos and a logo.
Click here to visit the Case Study Database and upload your case study.