Public testbeds are essential for replicable experiments and meaningful comparisons on shared physical infrastructure. While many testbeds exist for battery-powered Internet of Things (IoT) systems, there is a lack of public testbeds for observing and profiling the distributed operation of energy-harvesting IoT systems, including battery-free devices. We fill this gap and present Shepherd Nova, the first public testbed designed to support experiments under repeatable energy-harvesting conditions. Shepherd Nova uses field-recorded harvesting data to supply power to devices, consistently replicating real-world spatio-temporal energy availability across multiple experiments. Its virtual power source supports diverse ambient energy sources, harvesting circuitry, and energy storage devices. Moreover, Shepherd Nova provides services like general-purpose input/output (GPIO) tracing, power profiling, and serial output logging, all of which can run synchronously and with high resolution. Sub-microsecond synchronization enables precise correlation between these observations and emulated energy-harvesting conditions, offering unprecedented insights into distributed energy-harvesting IoT systems. In this paper, we describe Shepherd Nova's design, characterize its performance, and demonstrate its capabilities through controlled experiments and an example test case. To access the testbed, documentation as well as open-source harvesting data, hardware designs, and code, visit https://testbed.nes-lab.org/.

An important class of cyber-physical systems relies on multiple agents that jointly perform a task by coordinating their actions over a wireless network. Examples include self-driving cars in intelligent transportation and production robots in smart manufacturing. However, the scalability of existing control-over-wireless solutions is limited as they cannot resolve overload situations in which the communication demand exceeds the available bandwidth. This article presents a novel co-design of distributed control and wireless communication that overcomes this limitation by dynamically allocating the available bandwidth to agents with the greatest need to communicate. Experiments on a real cyber-physical testbed with 20 agents, each consisting of a low-power wireless embedded device and a cart-pole system, demonstrate that our solution achieves significantly better control performance under overload than the state of the art. We further prove that our co-design guarantees closed-loop stability for physical systems with stochastic linear time-invariant dynamics.

Many-to-all communication is a prerequisite for many applications and network services, including distributed control and data replication. However, current solutions do not meet the scalability and latency requirements of emerging applications. This paper presents Mixer, a many-to-all broadcast primitive for dynamic wireless mesh networks. Mixer integrates random linear network coding (RLNC) with synchronous transmissions and approaches the order-optimal scaling in the number of messages to be exchanged. To achieve an efficient operation in real networks, we design Mixer in response to the theory of RLNC and the characteristics of physical-layer capture. Our experiments demonstrate, for example, that Mixer outperforms the state of the art by up to 3.8x and provides a reliability greater than 99.99 % even at a node moving speed of 60 km/h.

Cyber-physical systems (CPS) use distributed feedback loops to control physical processes. Designing practical distributed CPS controllers often benefits from a logically centralized approach, where each node computes the control law locally based on global knowledge of the system state. We present Mixer, an all-to-all communication scheme that enables all nodes in a multi-hop low-power wireless network to exchange sizable packets with one another. Mixer's design integrates synchronous transmissions with random linear network coding, harnessing the broadcast nature of the wireless medium. Results from testbed experiments with an early Mixer prototype show that our design reduces latency by 1.1-2.6× for 16-96-byte packets compared with the state of the art, while providing a reliability above 99.9% in most settings we test.