This article demonstrates an unexplored relay operating mode in Bluetooth Low Energy (BLE)-based Wireless Sensor Networks, enabled by the link-layer synchronization inherent to BLE periodic advertisements. This synchronization allows relay nodes to avoid the power-hungry 100% duty primary-channel scanning that conventional passive BLE receivers rely on for reliable packet capture. To translate this insight into practice, we present the design and implementation of Dynamic Primary-Channel Scanning (DPCS), a practical strategy that disables primary-channel scanning after periodic synchronization and autonomously re-enables it upon synchronization loss, thereby allowing relay nodes to remain asleep for most of their lifetime while still forwarding data on scheduled secondary-channel receptions. We integrate DPCS into a multi-hop, chain-based data propagation framework for linear BLE-based WSNs, where sensor nodes themselves act as relays and forward concatenated data hop-by-hop toward a sink. Enabled by the innate properties of the periodic advertising mechanism, DPCS is directly deployable on commercial off-the-shelf (COTS) BLE 5.x devices as an application-level control logic using the provisions of the BLE stack. Experiments with nRF52832-based nodes in five-node chains demonstrate over 99% per-hop data reception reliability indoors across 50–60 m with modest retransmissions. In outdoor deployments, the per-hop reliability consistently remains above 99% without retransmissions across a 180–200 m span, with rapid automatic re-synchronization after interruptions. Power profiling shows consistently lower energy consumption than both continuous and duty-cycled primary-channel scanning baselines, with up to two to three orders of magnitude savings over continuous scanning at a sensing interval of 60 s. This chain-based analysis provides a baseline for extending the proposed periodic-advertising-based relay operation to more complex network topologies.

In Wireless Sensor Networks (WSNs) involving frequent transmission of large sensor data, packet collisions degrade the performance and add to the energy consumption of the sensor nodes since they require additional packet retransmissions to compensate data loss. In this paper, applications like livestock health monitoring systems have been targeted that involve frequent transmission of large sized motion data sampled along multiple axes. We choose low power non-connectable Bluetooth Low Energy (BLE) extended advertisements for quick transmission of large sensor data, and propose a non-connectable BLE advertising based polling algorithm for eliminating packet collision. Contrary to the existing polling-based schemes that are implemented at the link layer, this is the first work to propose a broadcast-based polling algorithm for the "application" layer. This facilitates its adaptability with commercially available BLE chips in practical deployments, since it runs on top of the standardized BLE link layer that does not implement carrier sensing. It has been shown through real-world experimentation carried out using nRF52832 and nRF52840 SoCs for 5 days, that the proposed polling technique is more effective in networks with higher transmission rates. In a network of 25 nodes transmitting 3,600 bytes every 30 s, with two packet re-transmissions, polling achieves a consistent data delivery success rate between 98-99%, as opposed to nearly 95% achieved without polling in higher (three) re-transmissions. Higher packet delivery rates are achievable in fewer re-transmissions with polling, thus saving power of the sensor nodes.

Bluetooth low energy (BLE) allows connectionless data transfer to take place through advertisements. Bluetooth Core Specification version 5.0 introduced extended advertising which allows each secondary channel advertising packet to carry up to 255 bytes of payload. Up to 1650 bytes of data can be advertised as a single transmission through chaining of multiple packets carrying fragments of advertising data. This article derives throughput expression in the presence of bit errors and analyses the maximum throughput for connectionless data transfer using extended advertisements, for all the Uncoded and Coded PHYs that offer different data rate and range capabilities. Guidelines are provided for the choice of appropriate PHYs based on the transmission size, throughput, and range requirement of the application. Simulation results have been presented that verify the analytically obtained throughput values. Since connectionless mode of communication does not allow acknowledgments from the receiver, it is essential to overcome the impact of bit errors through retransmission of each packet for increased chances of its reception. Analysis of maximum achievable throughput with retransmission of each data packet has also been presented in this article. Results suggest that the maximum achievable throughput is impacted by the duration of each transmission, due to which beyond a certain value of the data size, the low energy (LE) Coded PHYs are unable to offer significant improvement in the amount of data delivered per unit time.

In this paper, design of a BLE periodic advertising based data logger for cold chain temperature monitoring has been proposed, which is implemented using multi-threading on a Real Time Operating System for smooth, unhindered operation. The low-power logger design only implements the data fetching, storing and transmission operations, and the logged data can be accessed on a smartphone app using BLE periodic advertisements, through a simple button press, as many times as required to achieve 100% reception in the app. The data processing capabilities are implemented in the smartphone app and not the device to save power. Data transmission consumes ultra low current by the use of BLE periodic advertisements, which are experimentally shown in this paper to consume an average current of only 99.3 µA for transmitting the data logged over 24 hours. Additionally, even without entering into a connection, the data reception using BLE periodic advertising has been experimentally found to be in the range 96–98 % due to its ability to synchronize with the receiver, and the use of secondary advertising channels for data transfer. We have developed a smartphone app that implements a BLE periodic scanner, and allows the user to successfully receive 100% of the logged data, save it to the device/cloud and export the data anomalies based on the ambient range of values provided.

Modbus protocol is widely used in industrial automation, that allows the transfer of sensor data to take place over a serial interface, using a request-response framework. However, the deployment of Modbus relies on cabling and its scalability is restricted. In this paper, we propose an industrial automation framework that uses the low power connection-less mode of Bluetooth Low Energy to transfer data from the sensor nodes to a central Gateway using extended advertisements, and uses Modbus to relay the sensor data to the main supervisory unit. The Gateway stores all the packets received from the end nodes in a local storage, which is accessed each time a Modbus request for the data of a particular node is received. The paper proposes a customized wired data exchange algorithm developed on top of the standard Modbus RTU protocol, and efficient memory access (read/write) algorithms that ensure quick fetching of the sensor data from the storage at the Gateway corresponding to each Modbus request, making the response time independent of the size of the sensor network. We have provided real-time experimental results to show the high reception rates (94-100%) of sensor data transmitted using BLE extended advertisements in dense deployments, quick responsiveness of the Gateway (almost constant request-response time of 250 ms approximately), and the successful reception of sensor data at the supervisory unit using the proposed framework.

IoT applications like livestock and personalized health care monitoring systems have large amounts of data obtained per second from sensors interfaced to battery operated sensor nodes. In such networks, all the data generated at the nodes must be successfully sent to the central receiving device ensuring minimal power consumption of the data transmission operation, to enable longer lifetimes of battery-operated sensor nodes. In this article, in order to minimize energy consumption of sensor nodes, we propose a Bluetooth low energy (BLE)-based data transfer approach relying on BLE periodic advertisements, which allow up to 255 bytes of payload to be sent in one packet and minimize the involvement of interference prone primary advertising channels in the data transfer process. By using primary channel transmissions only for one-time device discovery and then disabling them, the proposed technique ensures that number of transmissions made by each sensor node for successful data delivery in large sensor networks is reduced in comparison with conventional extended and periodic advertisements. This offers reduction in energy consumption of sensor nodes due to reduction in radio on time. Through extensive real-time and simulation-based experimentation conducted, we have shown that, in a network of 200 sensor nodes, the proposed periodic advertising-based technique requires much lesser radio on time for reliable data delivery and reduces energy consumption of the sensor nodes by almost 51% and 20% as compared to the consumption when extended and standardized periodic advertisements are used to communicate the data, respectively.

Bluetooth Low energy (BLE) advertisements have great potential in Internet of Things applications like monitoring systems, which involve the real-time transfer of data collected from sensors interfaced to multiple power-constrained devices. Since sensor data conveys information about the present state of the system, packet loss is vital in such applications. The paper proposes a customized BLE legacy advertising packet structure to advertise dynamic sensor data, and proposes two techniques: One-Set-Multiple-Events (OSME) and One-Set-X-Events (OSXE), to advertise the dynamic sensor data using the non-connectable and non-scannable undirected advertising mode of BLE. Using appropriate advertising parameters with OSXE, each advertising packet gets successfully advertised$X$times, whereas OSME neither guarantees successful advertisement of all the packets, nor does it allow control over the number of transmitted copies of each packet. Using OSXE, experimental analysis is carried out to study packet loss when single BLE node transmits alone, and when 25 nodes transmit simultaneously. Mathematical analysis has been carried out for packet collision for multiple nodes. Experimental results show that packet loss is also impacted by factors like advertising set duration, scan window, scanner's link layer activities like advertising report generation and channel switching. Using OSXE, packet loss close to 0% is obtained for$X=3$when 25 nodes advertise simultaneously.

Internet-of-Things (IoT) applications require a network that covers a large geographic area, consumes less power, is low-cost, and is scalable with an increasing number of connected devices. Low-power wide-area networks (LPWANs) have recently received significant attention to meet these requirements of IoT applications. Long-range wide-area network (LoRaWAN) with long range (LoRa) (the physical layer design for LoRaWAN) has emerged as a leading LPWAN solution for IoT. However, LoRa networks suffer from the scalability issue when supporting a large number of end devices that access the shared channels randomly. The scalability of LoRa networks greatly depends on the spreading factor (SF) allocation schemes. In this article, we propose an exponential windowing scheme (EWS) for LoRa networks to improve the scalability of LoRa networks. EWS is a distance-based SF allocation scheme. It assigns a distance parameter to each SF to maximize the success probability of the overall LoRa network. Using stochastic geometry, expressions for success probability are derived under co-SF interference. The impact of exponential windowing and packet size is analyzed on packet success probability. In addition, the proposed scheme is compared with the existing distance-based SF allocation schemes: equal-interval-based and equal-area-based schemes, and it is shown that the proposed scheme performs better than the other two schemes.