Kazuya Echigo

Bio: Kazuya Echigo is an academic researcher from University of Tokyo. The author has contributed to research in topics: Mathematical optimization & Router. The author has an hindex of 2, co-authored 2 publications receiving 12 citations.

Detecting Presence From a WiFi Router’s Electric Power Consumption by Machine Learning

Abstract: Presence and occupancy detection in residential and office environments is used to predict movement of people, detect intruders, and manage electric power consumption. Specifically, we are developing methods to improve demand side electrical power management by reducing electrical power waste in unoccupied spaces. In this paper, we conduct an extensive analysis on the applicability of using a WiFi router’s electrical power consumption in different types of environments to determinate the number or people present in a space. We show the importance of a moving average filter for electrical load time series data, confirm the correlation between control packets and increased minimal router power consumption, and present our results on the accuracy of our approach. We conclude that a WiFi router’s power consumption can improve presence detection in home environments and occupancy estimation in office environments, and where possible, should be analysed separately from the aggregated power consumption.

Determining occupancy from a WiFi router's electric power consumption in an office environment

Abstract: The goal of most modern energy management systems in buildings is to reduce unnecessary waste and cut peak electric demand. Underpinning this need is the ability to detect the number of people occupying a given space. This allows for both the ability to shut down all nonessential devices contributing to the electric waste while no one is present, as well as optimise the amount of electricity used in heating, ventilation and air conditioning, depending on the number of people present. In this paper we demonstrate that it is possible to determine the number of people in a room by monitoring the electric power consumption of a WiFi router. We show that the WiFi router's minimum power consuption can be used as features to determine the number of connected devices and link this to the types of data packets used by the network. Finally, we also perform measurements in an office environment and show that it is possible to classify the number of people with high accuracy.

Linear Programming Approach to Relative-Orbit Control with Element-wise Quantized Control

Abstract: This paper considers the problem of close-proximity relative-orbit control of spacecraft with control quantization constraints, models the problem as a mixed-integer program, and reformulates the problem as a linear program. The reformulation uses linearized relative orbital element dynamics with a sum-of-absolute-values objective, and it is proved that optimal controls for the reformulated problem satisfy the quantization constraint, that is, the quantization constraints are convexified for the reformulated continuous-time optimal control problem. This problem is then discretized, converted to a finite-dimensional linear program, and solved using commercially available convex optimization solvers with polynomial-time convergence guarantees. Since the mathematical proofs of convexification are available only for continuous-time case, their validity for discrete-time is demonstrated with extensive simulations. To this end, Monte Carlo simulations indicate that quantization is achieved with high probability while keeping spacecraft slew rates low relative to other proposed approaches.

Convex Trajectory Planning for Proximity Operations using Electric Propulsion with Quantized Thrust

Abstract: This paper presents a convex optimization-based trajectory planner for close proximity relative-orbit operations using electric propulsion (EP). We target an EP system where the magnitude of thrust should be quantized. Traditional solution methods dealing with this input quantization-type constraint require solving computationally expensive mixed-integer programs. We propose an objective function that promotes quantization for the magnitude of the control input which, when combined with lossless convexifacation, results in a far more computationally tractable Second Order Cone Program (SOCP). A convex optimization problem minimizing this objective function is then formulated while satisfying a modified form of linearized relative orbital element (LROE) dynamics, control constraints, and boundary conditions. Additionally, we introduce a fuel sub-optimal constraint and propose a planner that relaxes the constraint recursively until a quantized solution is achieved. We demonstrate the capability of the proposed objective function through several numerical studies and provide a heuristic for achieving a quantized thrust profile.

[ACM Press the 2nd ACM Workshop - Zurich, Switzerland (2010.11.02-2010.11.02)] Proceedings of the 2nd ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Building - BuildSys \'10 - Occupancy-driven energy management for smart building automation

Abstract: Buildings are among the largest consumers of electricity in the US. A significant portion of this energy use in buildings can be attributed to HVAC systems used to maintain comfort for occupants. In most cases these building HVAC systems run on fixed schedules and do not employ any fine grained control based on detailed occupancy information. In this paper we present the design and implementation of a presence sensor platform that can be used for accurate occupancy detection at the level of individual offices. Our presence sensor is low-cost, wireless, and incrementally deployable within existing buildings. Using a pilot deployment of our system across ten offices over a two week period we identify significant opportunities for energy savings due to periods of vacancy. Our energy measurements show that our presence node has an estimated battery lifetime of over five years, while detecting occupancy accurately. Furthermore, using a building simulation framework and the occupancy information from our testbed, we show potential energy savings from 10% to 15% using our system.

A Reinforcement Learning-Based QAM/PSK Symbol Synchronizer

Abstract: Machine Learning (ML) based on supervised and unsupervised learning models has been recently applied in the telecommunication field. However, such techniques rely on application-specific large datasets and the performance deteriorates if the statistics of the inference data changes over time. Reinforcement Learning (RL) is a solution to these issues because it is able to adapt its behavior to the changing statistics of the input data. In this work, we propose the design of an RL Agent able to learn the behavior of a Timing Recovery Loop (TRL) through the Q-Learning algorithm. The Agent is compatible with popular PSK and QAM formats. We validated the RL synchronizer by comparing it to the Mueller and Muller TRL in terms of Modulation Error Ratio (MER) in a noisy channel scenario. The results show a good trade-off in terms of MER performance. The RL based synchronizer loses less than 1 dB of MER with respect to the conventional one but it is able to adapt its behavior to different modulation formats without the need of any tuning for the system parameters.

Honeypot Type Selection Games for Smart Grid Networks

Abstract: In this paper, we define a cyber deception game between the Advanced Metering Infrastructure (AMI) network administrator (henceforth, defender) and attacker. The defender decides to install between a low-interaction honeypot, high-interaction honeypot, and a real system with no honeypot. The attacker decides on whether or not to attack the system given her belief about the type of device she is facing. We model this interaction as a Bayesian game with complete but imperfect information. The choice of honeypot type is private information and characterizes the essence and objective of the defender i.e., the degree of deception and amount of threat intelligence. We study the players’ equilibrium strategies and provide numerical illustrations. The work presented in this paper has been motivated by the H2020 SPEAR project which investigates the implementation of honeypots in smart grid infrastructures to: (i) contribute towards creating attack data sets for training a SIEM (Security Information and Event Management) and (ii) to support post-incident forensics analysis by having recorded a collection of evidence regarding an attacker’s actions.

Detecting Presence From a WiFi Router’s Electric Power Consumption by Machine Learning

Abstract: Presence and occupancy detection in residential and office environments is used to predict movement of people, detect intruders, and manage electric power consumption. Specifically, we are developing methods to improve demand side electrical power management by reducing electrical power waste in unoccupied spaces. In this paper, we conduct an extensive analysis on the applicability of using a WiFi router’s electrical power consumption in different types of environments to determinate the number or people present in a space. We show the importance of a moving average filter for electrical load time series data, confirm the correlation between control packets and increased minimal router power consumption, and present our results on the accuracy of our approach. We conclude that a WiFi router’s power consumption can improve presence detection in home environments and occupancy estimation in office environments, and where possible, should be analysed separately from the aggregated power consumption.