Search for: All records

Inertial navigation provides a small footprint, low-power, and low-cost pathway for localization in GPS-denied environments on extremely resource-constrained Internet-of-Things (IoT) platforms. Traditionally, application-specific heuristics and physics-based kinematic models are used to mitigate the curse of drift in inertial odometry. These techniques, albeit lightweight, fail to handle domain shifts and environmental non-linearities. Recently, deep neural-inertial sequence learning has shown superior odometric resolution in capturing non-linear motion dynamics without human knowledge over heuristic-based methods. These AI-based techniques are data-hungry, suffer from excessive resource usage, and cannot guarantee following the underlying system physics. This paper highlights the unique methods, opportunities, and challenges in porting real-time AI-enhanced inertial navigation algorithms onto IoT platforms. First, we discuss how platform-aware neural architecture search coupled with ultra-lightweight model backbones can yield neural-inertial odometry models that are 31–134 x smaller yet achieve or exceed the localization resolution of state-of-the-art AI-enhanced techniques. The framework can generate models suitable for locating humans, animals, underwater sensors, aerial vehicles, and precision robots. Next, we showcase how techniques from neurosymbolic AI can yield physics-informed and interpretable neural-inertial navigation models. Afterward, we present opportunities for fine-tuning pre-trained odometry models in a new domain with as little as 1 minute of labeled data, while discussing inexpensive data collection and labeling techniques. Finally, we identify several open research challenges that demand careful consideration moving forward. 

Robotically assisted painting is widely used for spray and dip applications. However, use of robots for coating substrates using a roller applicator has not been systematically investigated. We showed for the first time, a generic robot arm-supported approach to painting engineering substrates using a roller with a constant force at an accurate joint step, while retaining compliance and thus safety. We optimized the robot design such that it is able to coat the substrate using a roller with a performance equivalent to that of a human applicator. To achieve this, we optimized the force, frequency of adjustment, and position control parameters of robotic design. A framework for autonomous coating is available athttps://github.com/duyayun/Vision-and-force-control-automonous-painting-with-rollers; users are only required to provide the boundary coordinates of surfaces to be coated. We found that robotically- and human-painted panels showed similar trends in dry film thickness, coating hardness, flexibility, impact resistance, and microscopic properties. Color profile analysis of the coated panels showed non-significant difference in color scheme and is acceptable for architectural paints. Overall, this work shows the potential of robot-assisted coating strategy using roller applicator. This could be a viable option for hazardous area coating, high-altitude architectural paints, germs sanitization, and accelerated household applications.

 

Precision agricultural robots require high-resolution navigation solutions. In this paper, we introduce a robust neural-inertial sequence learning approach to track such robots with ultra-intermittent GNSS updates. First, we propose an ultra-lightweight neural-Kalman filter that can track agricultural robots within 1.4 m (1.4–5.8× better than competing techniques), while tracking within 2.75 m with 20 mins of GPS outage. Second, we introduce a user-friendly video-processing toolbox to generate high-resolution (±5 cm) position data for fine-tuning pre-trained neural-inertial models in the field. Third, we introduce the first and largest (6.5 hours, 4.5 km, 3 phases) public neural-inertial navigation dataset for precision agricultural robots. The dataset, toolbox, and code are available at: https://github.com/nesl/agrobot. 

Abstract Flagella and cilia are slender structures that serve important functionalities in the microscopic world through their locomotion induced by fluid and structure interaction. With recent developments in microscopy, fabrication, biology, and modeling capability, robots inspired by the locomotion of these organelles in low Reynolds number flow have been manufactured and tested on the micro-and macro-scale, ranging from medical in vivo microbots, microfluidics to macro prototypes. We present a collection of modeling theories, control principles, and fabrication methods for flagellated and ciliary robots.