18 July 2018 in IEEE Transactions on Vehicular Technology
In narrow band-Internet of Things (NB-IoT) systems, a preamble of physical random access channel (NPRACH) is based on a single-tone waveform combined with frequency hopping, which is appropriate to achieve the goals of IoT, such as long battery lifetime and extended coverage. Since the hopping pattern of a preamble is predetermined, it is possible to estimate time-of-arrival (ToA) using phase difference. The accurate ToA is very important for device positioning and uplink signal decoding. In this paper, we design a new frequency hopping pattern of NPRACH preamble which fully utilizes all feasible hopping distances for a given number of subcarriers in NPRACH. The proposed hopping pattern maintains the compatibility with the standard system, by keeping the NPRACH structure in standard with minimal change. The simulation results show that the proposed hopping pattern greatly improves the accuracy of ToA estimation without additional system overhead.
8 April 2020 in IEEE Access
IEEE 802.11-based wireless local area network (WLAN) is currently the most popular communication system and a lot of access points (APs) have been densely deployed over heavily populated areas. In such dense WLAN environment, the WLAN users can suffer from performance degradation due to high co-channel and adjacent-channel interference among APs. To tackling such user performance degradation problem, we propose a two-phase radio resource management (RRM) framework, of which the first phase is channel assignment (CA) and the second phase is user association for channel load balancing (UA-ChLB). In designing the RRM framework, we take account of typical WLAN environment where dual bands are supported and two types of APs coexist: controlled-APs, which are managed by a centralized controller, and stand-alone APs, which independently operate for themselves. The proposed CA method utilizes a channel bonding technique, which provides a high data rate by integrating multiple basic channels while reducing the interference among neighboring APs. The proposed hybrid UA-ChLB scheme, which is composed of distributed/centralized UA-ChLB, efficiently coordinates the channel load among neighboring APs and between two bands of 2:4 and 5 GHz, under consideration of the wireless channel quality, interference, and traffic conditions. We implement a prototype system in practice using the proposed RRM scheme. The experimental results show that the proposed RRM scheme greatly improves both throughput and fairness in dense WLAN environment, as compared with some existing schemes.
20 April 2021 in IEEE Internet of Things Journal
Recently, many researchers and global vendors of mobile devices have paid attention to the ranging technique using IEEE 802.15.4z ultra-wideband (UWB), because it can support a highly precise spatial awareness and indoor localization. In the Internet of Things (IoT) environments, the interval of the ranging attempts for a ranging device (RDEV) affects largely the performance of the RDEV. The proper ranging interval for an RDEV varies according to the IoT applications using the ranging results as well as the type and state of the RDEV (e.g., moving speed and battery status). In the IEEE 802.15.4z, since the ranging for a group of multiple RDEVs is carried out once every ranging block, the ranging interval can be represented by the block length, which should be determined based on the various requirements of RDEVs within the group. In this article, we analyze the requirements on the block length of RDEVs, from the viewpoint of the energy efficiency and the degree of service satisfaction. Considering these requirements, we design a novel decision scheme of the ranging block length, based on the Markov decision process. The simulation results show that the proposed scheme enhances remarkably the energy efficiency and the satisfaction of RDEVs.
8 April 2020 in IEEE Internet of Things Journal
Long Range wide area networks (LoRaWANs) have been designed to provide low-power long-range communication service to IoT applications. To efficiently support a massive number of IoT devices in wide area with limited resource, the LoRaWAN utilizes various adaptation techniques including rate and power control. In this paper, we propose a very simple and energy-efficient rate control scheme for LoRa uplink transmission. With the proposed scheme, a LoRa end device increases its rate (rate_up) when recognizing that the uplink channel quality gets better and, when recognizing transmission failure under uncongested network situation, the device decreases the rate (rate_down) or increases transmission power. The proposed scheme also has a simple mechanism for avoiding the repeated alternation between rate_up and rate_down, called a ping-pong situation. We evaluate the performance of proposed scheme, by modeling it as a continuous time Markov chain. In addition, we compare the performances of proposed scheme and some existing schemes by using simulation. The simulation results demonstrate that the proposed scheme greatly outperforms the others, in transmission success probability, effective transmission rate, frame transmission delay, and energy consumption.
1 April 2025 in IEEE Internet of Things Journal
Gesture recognition can provide a natural and user-friendly interface to a variety of systems. However, in order for the real-world applications to apply gesture recognition as a tool for human-computer interaction (HCI), it is necessary not only to accurately classify predefined gestures but also to correctly identify undefined gestures. In this paper, we design a novel open-set gesture recognition (OSR) system which leverages ultra-wideband (UWB) impulse radar-based wireless sensing and deep learning technology to effectively detect both known and unknown gestures. In the proposed system, the OSR task is decomposed to two steps of closed-set classification and deep inspection. The first closed-set classification step extracts the features from channel impulse response (CIR) matrix of UWB radar and predicts one of the known gestures. The second deep inspection step investigates more deeply whether the predicted known gesture is correct or an unknown gesture. The feature extractor is composed of two Transformers, which extract temporal features and spatial features unique to each gesture from the CIR matrix, respectively. The second stage consists of multiple binary deep inspectors, each of which is assigned to one of the known gestures, and this method not only greatly simplifies the training process of deep inspection but also remarkably enhances the OSR performance. Extensive experimental results conducted on a public dataset demonstrate that our system significantly outperforms the state-of-the-art OSR systems.
15 May 2019 in IEEE Systems Journal
As the IEEE 802.11 (WiFi) becomes the de-facto global standard for wireless local area network (WLAN), a huge number of WiFi access points (APs) are deployed and the mobile users enjoy almost ubiquitous wireless Internet access. When densely deployed APs covering a certain area are managed by a single organization (say, an enterprise), there can exist multiple candidate APs for serving a user. In this case, the user association (UA), i.e., the selection of serving AP, holds a key role in the network performance. In this paper, we propose an efficient UA scheme for load balancing among APs and energy saving in the enterprise WLAN, considering various techniques for performance improvement, such as multicast transmission, multiuser multiple-input and multiple-output, and AP sleeping, together. We formulate the problem into a multiobjective optimization and get the solution as the UA scheme. We develop a WLAN testbed to demonstrate the practicality and feasibility of the proposed scheme. Moreover, through extensive simulations, we also compare the performance of the proposed scheme with those of some other UA schemes. The simulation results show that the proposed scheme outperforms the other schemes in terms of the total system throughput, the throughput per working AP, the energy efficiency, and the load balancing factor.