Davide Del Testa

TL;DR: A convolutional neural network is trained to map raw pixels from a single front-facing camera directly to steering commands and it is argued that this will eventually lead to better performance and smaller systems.

Abstract: We consider an energy harvesting device whose state at a given time is determined by its energy level and an “importance” value, associated to the transmission of a data packet to the receiver at that particular time. We consider policies that, at each time, elect whether to transmit the data packet or not, based on the current energy level and data importance, so as to maximize the long-term average transmitted data importance. Under the assumption of i.i.d. Bernoulli energy arrivals, we show that the sensor should report only data with an importance value above a given threshold, which is a strictly decreasing function of the energy level, and we derive upper and lower bounds on the thresholds for any energy level. Leveraging on these findings, we construct a suboptimal policy that performs very close to the optimal one, at a fraction of the complexity. Finally, we demonstrate that a threshold policy, which on the average transmits with probability equal to the average energy arrival rate is asymptotically optimal as the energy storage capacity grows large.

TL;DR: This work advocates the use of lossy signal compression as a means to decrease the data size of the gathered biosignals and, in turn, boost the battery life of wearables and allow for fine-grained and long-term monitoring.

TL;DR: This letter advocates the use of autoencoders as an efficient and computationally lightweight means to compress biometric signals, and applies them to ECG traces, showing quantitative results in terms of compression ratio, reconstruction error and computational complexity.

TL;DR: This paper considers a wireless network composed of a pair of sensors powered by energy harvesting devices (EHDs), which transmit data to a receiver over a shared wireless channel, and derives the optimal policy in the cases where the state of charge (SOC) may not be perfectly known by the central controller.