Disclaimer: These are sample project descriptions submitted by participating faculty. The actual projects pursued in Summer 2017 may be the same or change. Final project assignments will be finalized after student acceptance to the program. Some projects will have more than one student assigned to them.
- Bioinspired robotics in space
- Insect inspired microrobots
- Multifunctional skins for bioinspired robots
- Bioinspired robot birds
- Navigation for bioinspired robotics
- Dynamics and controls of soft robots on land and in water
- Actuators for bioinspired robots
- Bioinspired sensors for robotics
The Space Systems Lab works on numerous projects related to space robotics and human-robot interaction. The environments in which robots are used in space applications are quite challenging, therefore this project will focus on bio-inspired space robotics for tasks from satellite capture and repair to rovers. Students will have the opportunity to develop new bio-ispired robots and validate behavior in space-simulated environments such as our novel Neutral Buoyancy facility.
Insects display amazing locomotion capabilities at high speeds over rough surfaces strewn with obstacles. If microrobots could move in a similar fashion, they would be able to climb through rubble to find survivors after an earthquake or build impressive structures similar to ant and termite mounds. However, achieving the same feats with a microrobot less than 1 cm in size is not a trivial problem. As part of a project to improve the efficacy and efficiency of microrobot locomotion, REU students will investigate new robot designs through both analysis and experimental testing that can take advantage of features such as springy legs (to return energy) or added damping (to improve stability over rough terrain). This project will allow students to learn more about the mechanics and dynamics of a crawling insect-inspired robot in addition to new techniques used to fabricate robots only centimeters in size or smaller.
To realize bioinspired robotic platforms that have the same capabilities as biological counterparts, it is necessary to have a paradigm shift in the types of structures that are used. In particular, it is important to create multifunctional skins, which would be an assembly of sensors, actuators, power supplies, and structural components. This project will enable REU students to investigate and characterize new materials that could provide multifunctional elastomeric structures or structures that could also be used to store energy for the robot's future consumption. For example, multifunctional skins that can harvest and store solar energy to enhance the operational time of a robot for more autonomy. Students will learn about fundamental principles for modeling and designing materials and structures based on actuation, sensing, and power systems in bioinspired robots through this project. They will also learn about new methods for creating this skins using techniques such as additive manufacturing.
Successful realization of a flapping wing micro air vehicle (MAV) requires development of a light weight drive mechanism converting the rotary motion of the motor into flapping motion of the wings. In order to make flapping wing MAVs attractive in search, rescue, and recovery missions, they should be disposable from the cost point of view. For example, we have developed a novel flapping wing MAV that is bird-inspired known as Robo Raven [see video with over 260,000 views at https://www.youtube.com/watch?v=mjOWpwbnmTw]. It has servo motors that independently control each wing enabling it to perform aerobatic maneuvers like real birds. It is so realistic, it was caught on video being attacked by a hawk and leading a flock of birds. As a part of this project students will have an opportunity to develop and understand the physics and associated control algorithms enabling wings to change their position and speed instantaneously in order to perform maneuvers autonomously, such as controlled dives and loitering. The student will also learn and apply kinematics and dynamics principles during this project that are essential to modeling the forces that control the aerobatic maneuvers.
We are involved in the design, construction, and testing of many new sonar configurations and algorithms used to analyze the resulting sensory data to perform object localization. We are developing narrowband and broadband ultrasonic sonar sensors, sensors for 2-D and 3-D localization, and models of biologically-realistic and non-biological sonar systems. These will be tailored for both biological modeling and for general robotics development. Beyond the localization of single echo, we are developing models of multi-echo feature extraction to represent clusters of objects in the environment in task-relevant terms (e.g., obstacle, wall, corridor, complete blockage, recognition of objects, etc.) These projects are centered on neural models of sensory processing in the bat brainstem and midbrain. A major goal of research in the lab is development of neural models of spatial perception and navigation in echolocating bats. While neurobiological implementation of spatial reasoning is well beyond the scope of an undergraduate project, there are many appropriate sub-projects. For example, we are building models of biological odometry, models of mapping spaces with robots, and neurally-inspired models of path planning. Students will be exposed to many concepts from neurobiology, neural computation, machine learning, and robotics.
The fundamental questions to be addressed in the area of dynamics and control of a soft robotic system with distributed sensing and actuation include the following: to what extent can the natural dynamics of a soft, flexible structure be exploited in its closed-loop operation (e.g., to avoid using too much actuation and/or energy); and how might we apply tools from the control of networked systems to design a feedback control system for a soft robot with many spatially distributed sensors (e.g., mechanosensitive hairs covering surfaces) and actuators? Networked sensors and control that achieve synchronization and other global patterns using (only) local connections have the potential to advance the state-of-the-art in soft robotic locomotion, much as they have advanced the field of legged robotics. Students will have opportunities to participate in development of a novel soft robotic testbed to demonstrate robust undulatory motion in variable-friction terrain via dynamic control of bending and adhesion. They will also be able to explore closed-loop bio-sensing and bio-actuation for soft robots in water by developing small, fish-inspired underwater vehicles where compliant structures increase the available degrees of freedom to ultimately achieve free-swimming robots. Experimental facilities available to students include a flowing water tank, a water tow tank, and UMD's Neutral Buoyancy Research Facility.
To realize the proposed robotic platforms, it would be beneficial to have new types of actuators. The ability to create compliant actuators with high force and stroke that can be scaled in size and shape for locomotion, gripping, and other tasks at relatively low power is therefore another focus. We have previously demonstrated "nastic" actuators based on hydraulics using electroosmotic flow (EOF) within channels in elastomeric substrates. These actuators have the potential to be both high force, since they are based on hydraulics, and high stroke, provided by elastomeric membranes. To improve the performance of the pumping, we have recently switched from water as the pumping fluid to the organic solvent propylene carbonate, which does not generate gases even at 10 kV. To increase force we are now using porous matrices rather than microchannels, and we are also investigating the use of carbon-based compliant electrodes. It would also be beneficial for these platforms to have built-in strain sensing. We have been developing piezoresistive sensors based on carbon-loaded elastomers, which can be spin-coated to integrate with other layers or painted to retrofit onto existing structures. For the proposed training activity, we will build upon our previous experience manufacturing compliant electrodes and strain gauges, focusing on the preparation and characterization of the composite materials as well as methods for forming robust and reliable electrical connections to the compliant conductors.
The survival of animals depends on their effectiveness in collecting sufficient and timely information about their ever-changing environment and on their ability to act upon sensory information. For mobile autonomous robots and vehicles (such as micro air vehicles (MAVs)), analogous capabilities are desirable, but far from being well developed. Specifically, bioinspired robots equipped with directional hearing and sound localization ability can locate objects within a full 360o field-of-view even in a dark environment (e.g., at night), which is a remarkable improvement over vision field-of-view that is restricted to be less than 180o. Researchers have found that the fly Ormia utilizes a unique localization-lateralization scheme for achieving superior sound localization precision. In the proposed undergraduate research project, students will study an off-the-shell small robot equipped with fly-ear inspired bioinspired directional microphones (previously developed by Prof. Yu's group) for acoustic localization, homing, and navigation. A simple, yet effective acoustic localization algorithm inspired by the fly's localization/lateralization scheme will be implemented and tested for robotic acoustic localization. Specific research tasks including sensor instrumentation, integration of sensors with robots, testing the implementation of the bioinspired algorithms, and characterization of the system performance for localization, homing, and navigation.