An official website of the United States government
Autonomous systems, such as Unmanned Aerial Vehicles (UAVs), are expected to run complex reinforcement learning (RL) models to execute fully autonomous positionnavigation-time tasks within stringent onboard weight and power constraints. We observe that reducing onboard operating voltage can benefit the energy efficiency of both the computation and flight mission, however, it can also result in on-chip bit failures that are detrimental to mission safety and performance. To this end, we propose BERRY, a robust learning framework to improve bit error robustness and energy efficiency for RL-enabled autonomous systems. BERRY supports robust learning, both offline and on-board the UAV, and for the first time, demonstrates the practicality of robust low-voltage operation on UAVs that leads to high energy savings in both compute-level operation and systemlevel quality-of-flight. We perform extensive experiments on 72 autonomous navigation scenarios and demonstrate that BERRY generalizes well across environments, UAVs, autonomy policies, operating voltages and fault patterns, and consistently improves robustness, efficiency and mission performance, achieving up to 15.62% reduction in flight energy, 18.51% increase in the number of successful missions, and 3.43× processing energy reduction.
more »
« less
Multi-modal jumping and crawling in an autonomous, springtail-inspired microrobot
Springtails are tiny arthropods that crawl and jump. They jump by temporarily storing elastic energy in resilin elastic cuticular structures and releasing that energy to accelerate a tail, called a furca, propelling them in the air. This paper presents an autonomous, springtail-inspired microrobot that can crawl and jump. The microrobot has a mass of 980mg and stands 13mm tall, and has on-board sensing, computation, and power, enabling autonomy. The microrobot was designed with a super-elastic shape memory alloy (SMA) spring that is manually loaded to store elastic energy. The on-board sensing and computation triggers an actuator at the jump frequency range that unlatches the spring, launching the microrobot into the air at speeds up to 3.171 m/s. At the same time, the microrobot is capable of crawling, when actuated at frequencies lower or higher than the jump frequency range, demonstrating autonomous multi-modal locomotion. This work opens up new pathways toward autonomy in multi-modal microrobots.
more »
« less
- Award ID(s):
- 2153327
- PAR ID:
- 10545224
- Publisher / Repository:
- IEEE
- Date Published:
- ISBN:
- 979-8-3503-8457-4
- Page Range / eLocation ID:
- 5999 to 6005
- Format(s):
- Medium: X
- Location:
- Yokohama, Japan
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Autonomous edge computing in robotics, smart cities, and autonomous vehicles relies on the seamless integration of sensing, processing, and actuation for real-time decision-making in dynamic environments. At its core is the sensing-to-action loop, which iteratively aligns sensor inputs with computational models to drive adaptive control strategies. These loops can adapt to hyper-local conditions, enhancing resource efficiency and responsiveness, but also face challenges such as resource constraints, synchronization delays in multimodal data fusion, and the risk of cascading errors in feedback loops. This article explores how proactive, context-aware sensing-to-action and action-to-sensing adaptations can enhance efficiency by dynamically adjusting sensing and computation based on task demands, such as sensing a very limited part of the environment and predicting the rest. By guiding sensing through control actions, action-to-sensing pathways can improve task relevance and resource use, but they also require robust monitoring to prevent cascading errors and maintain reliability. Multi-agent sensing-action loops further extend these capabilities through coordinated sensing and actions across distributed agents, optimizing resource use via collaboration. Additionally, neuromorphic computing, inspired by biological systems, provides an efficient framework for spike-based, event-driven processing that conserves energy, reduces latency, and supports hierarchical control-making it ideal for multi-agent optimization. This article highlights the importance of end-to-end co-design strategies that align algorithmic models with hardware and environmental dynamics, improve cross-layer inter-dependencies to improve throughput, precision, and adaptability for energy-efficient edge autonomy in complex environments.more » « less
-
Recent advances in precision manufacturing technology and a thorough understanding of the properties of piezoelectric materials have made it possible for researchers to develop innovative microrobotic systems, which draw more attention to the challenges of utilizing microrobots in areas that are inaccessible to ordinary robots. This review paper provides an overview of the recent advances in the application of piezoelectric materials in microrobots. The challenges of microrobots in the direction of autonomy are categorized into four sections: mechanisms, power, sensing, and control. In each section, innovative research ideas are presented to inspire researchers in their prospective microrobot designs according to specific applications. Novel mechanisms for the mobility of piezoelectric microrobots are reviewed and described. Additionally, as the piezoelectric micro-actuators require high-voltage electronics and onboard power supplies, we review ways of energy harvesting technology and lightweight micro-sensing mechanisms that contain piezoelectric devices to provide feedback, facilitating the use of control strategies to achieve the autonomous untethered movement of microrobots.more » « less
-
Millions of years of evolution have allowed animals to develop unusual locomotion capabilities. A striking example is the legless-jumping of click beetles and trap-jaw ants, which jump more than 10 times their body length. Their delicate musculoskeletal system amplifies their muscles’ power. It is challenging to engineer insect-scale jumpers that use onboard actuators for both elastic energy storage and power amplification. Typical jumpers require a combination of at least two actuator mechanisms for elastic energy storage and jump triggering, leading to complex designs having many parts. Here, we report the new concept of dynamic buckling cascading, in which a single unidirectional actuation stroke drives an elastic beam through a sequence of energy-storing buckling modes automatically followed by spontaneous impulsive snapping at a critical triggering threshold. Integrating this cascade in a robot enables jumping with unidirectional muscles and power amplification (JUMPA). These JUMPA systems use a single lightweight mechanism for energy storage and release with a mass of 1.6 g and 2 cm length and jump up to 0.9 m, 40 times their body length. They jump repeatedly by reengaging the latch and using coiled artificial muscles to restore elastic energy. The robots reach their performance limits guided by theoretical analysis of snap-through and momentum exchange during ground collision. These jumpers reach the energy densities typical of the best macroscale jumping robots, while also matching the rapid escape times of jumping insects, thus demonstrating the path toward future applications including proximity sensing, inspection, and search and rescue.more » « less
-
Abstract Jumping microrobots and insects power their impressive leaps through systems of springs and latches. Using springs and latches, rather than motors or muscles, as actuators to power jumps imposes new challenges on controlling the performance of the jump. In this paper, we show how tuning the motor and spring relative to one another in a torque reversal latch can lead to an ability to control jump output, producing either tuneable (variable) or stereotyped jumps. We develop and utilize a simple mathematical model to explore the underlying design, dynamics, and control of a torque reversal mechanism, provides the opportunity to achieve different outcomes through the interaction between geometry, spring properties, and motor voltage. We relate system design and control parameters to performance to guide the design of torque reversal mechanisms for either variable or stereotyped jump performance. We then build a small (356 mg) microrobot and characterize the constituent components (e.g. motor and spring). Through tuning the actuator and spring relative to the geometry of the torque reversal mechanism, we demonstrate that we can achieve jumping microrobots that both jump with different take-off velocities given the actuator input (variable jumping), and those that jump with nearly the same take-off velocity with actuator input (stereotyped jumping). The coupling between spring characteristics and geometry in this system has benefits for resource-limited microrobots, and our work highlights design combinations that have synergistic impacts on output, compared to others that constrain it. This work will guide new design principles for enabling control in resource-limited jumping microrobots.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/ICRA57147.2024.10610130
