A Hybrid Anti-Collision Protocol Based on Frequency Division Multiple Access (FDMA) and Time Division Multiple Access (TDMA) for Radio Frequency Identification (RFID) Readers

Abstract

Radio Frequency Identification (RFID) technology plays a crucial role in various Internet of Things (IoT) applications, necessitating the integration of RFID systems into dense networks. However, the presence of numerous readers leads to collisions, degrading communication between readers and tags and compromising system performance. To tackle this challenge, researchers have proposed Medium Access Control (MAC) layer protocols employing different channel access methods. In this paper, we present a novel solution, the Distributed Time Slot Anti-Collision protocol (DTS-AC), which employs a new TDMA notification system to address Reader-to-Reader Interference (RRI), while incorporating FDMA-based frequency resource management to resolve Reader-to-Tag Interference (RTI) collision issues. Simulation results demonstrate that DTS-AC significantly improves performance in dense RFID networks by enhancing read rates, with scalability benefits based on the number of readers, channels, and Time Slots (TSs). Moreover, the cost-effectiveness of DTS-AC facilitates efficient deployment in RFID networks, emphasizing considerations of time delay and data sensitivity.

1. Introduction

RFID, short for Radio Frequency Identification, is a method for remote data storage and collection using metal tags known as “RFID tags”. These tags respond to radio waves, enabling wireless transmission of information and showing promise in potentially replacing traditional barcodes [1]. Despite its efficiency, RFID adoption raises significant ethical and privacy concerns.

The versatility of RFID spans various sectors, including industry [2], commerce [3], healthcare [4], warehouses [5], tracking [6], and agriculture [7]. Central to an RFID system is its straightforward architecture, consisting primarily of two components: the reader and the tag [8,9].

In parallel with RFID technology, wireless communication networks have evolved rapidly, becoming integral to modern network architectures. Wireless Sensor Networks (WSNs) represent a significant application of wireless communication, offering seamless connectivity and communication. The integration of RFID with WSNs has been explored in several papers [10,11,12,13,14,15], addressing challenges related to energy consumption, convergence, and collision within network systems.

RFID systems operate across different frequency bands and bandwidths, each suited to specific ranges. For example, in retail, electronic price tags rely on RFID readers for accurate inventory management, while in warehouses, readers cover expansive areas for efficient tracking. To address coverage limitations, multiple RFID readers are deployed in specific domains, such as entry points in malls or checkpoints in warehouses. This tailored approach ensures comprehensive coverage, optimizing operational efficiency in various applications.

In densely populated RFID systems, one critical consideration is the bandwidth utilized. This selection is pivotal for optimizing system performance and minimizing interference, particularly in scenarios where multiple readers coexist within close proximity. For instance, low-frequency RFID may not be suitable for high-density concurrent requests, as its limited bandwidth can lead to increased collisions and reduced efficiency in densely populated environments. Collisions at the MAC layer are common in such environments, primarily involving readers.

In dense multi-channel RFID networks, Reader-to-Reader Interference (RRI) and Reader-to-Tag Interference (RTI) manifest in unique ways due to the presence of multiple communication channels. The definition of RRI and RTI in this context is as follows [16,17,18]:

-

Reader-to-Reader Interference (RRI) [Figure 1a]: In dense multi-channel RFID networks, RRI occurs when one RFID reader’s transmission overlaps with the transmission range (Discontinuous circle in Figure 1) of another nearby reader operating on a different channel. This interference disrupts the communication between the readers and the tags within their respective interrogation zones. Unlike in single-channel networks, where RRI primarily occurs due to overlapping transmission ranges, in dense multi-channel networks, RRI can also arise from interference caused by readers operating on adjacent or overlapping frequency bands. This can lead to collisions between reader signals, resulting in decreased read accuracy, increased latency, and reduced overall system performance.

-

Reader-to-Tag Interference (RTI) [Figure 1b]: RTI in dense multi-channel RFID networks occurs when multiple readers simultaneously attempt to interrogate the same set of tags, leading to interference at the tag level. This interference arises when tags within the reading range. (Continuous circle in Figure 1) of multiple readers receive conflicting signals, making it challenging for them to respond accurately to the interrogation commands. In dense multi-channel networks, RTI may be exacerbated by the presence of multiple readers operating on different channels, increasing the likelihood of tag collisions and communication errors. Consequently, RTI can lead to decreased tag read rates, increased collision probability, and reduced system efficiency.

Figure 1. RFID collisions. (a) Reader-to-Reader Interference; (b) Reader-to-Tag Interference.

Figure 1. RFID collisions. (a) Reader-to-Reader Interference; (b) Reader-to-Tag Interference.

Concurrent requests, the primary cause of reader collisions in RFID systems, occur when multiple readers contend for access to the communication channel concurrently. This phenomenon may stem from various environmental conditions, such as high reader density, reader proximity, and system congestion.

Understanding the implications of concurrent requests is crucial for developing effective collision mitigation strategies. By delving deeper into the causes and impacts of collisions in RFID systems, researchers and practitioners can devise more robust and efficient solutions to address these challenges.

In our paper, we introduce the Distributed Time Slot Anti-Collision protocol (DTS-AC), a novel solution designed to tackle the challenges of Reader-to-Reader Interference (RRI) and Reader-to-Tag Interference (RTI) collisions in RFID systems. Through a combination of innovative techniques, DTS-AC not only resolves these issues but also maximizes the number of active readers while minimizing reader collisions and reading delay.

1.1. Addressing RRI and RTI Collisions

DTS-AC employs a new TDMA notification system to mitigate RTI by allocating specific time slots to each reader, ensuring that they can operate independently without interfering with one another.

Additionally, by incorporating FDMA-based frequency resource management, DTS-AC effectively resolves RRI collisions. This approach assigns different frequency channels to individual readers, minimizing collisions during tag interrogation.

1.2. Maximizing Active Readers

A key feature of DTS-AC is its utilization of dual notification systems for frequency and timeslot management. This innovative approach allows for efficient allocation of resources, maximizing the number of active readers within the system.

Furthermore, DTS-AC implements a frequency reuse strategy, optimizing the utilization of available frequencies to accommodate a larger number of readers while ensuring efficient tag interrogation with minimal collisions.

1.3. Minimizing Reader Collisions and Reading Delay

DTS-AC prioritizes the minimization of reader collisions and reading delay to enhance network performance. By focusing on efficient resource allocation and management, DTS-AC minimizes the likelihood of reader collisions, leading to smoother operation and improved read rates.

Additionally, the protocol’s emphasis on minimizing tag reading delay is achieved through features such as Time Slot sharing among multiple readers, enabling faster and more efficient tag identification processes.

Through comprehensive simulations, we demonstrate that DTS-AC significantly improves the performance of RFID systems, particularly in dense network environments. Its scalability benefits, based on the number of readers, channels, and Time Slots, make it a promising solution for various RFID applications.

The remainder of this paper is structured as follows: Section 2 presents our proposed DTS-AC algorithm, Section 3 discusses related literature protocols, Section 4 presents the results of our simulation, and finally, Section 5 concludes the paper.

2. Related Work

In order to address collision problems, readers employ coordination mechanisms based on various communication architectures. These architectures can either be distributed among readers or be centralized under the control of a server. The choice of communication architecture determines the nature of the protocol and is influenced by the constraints of the environment.

The selection of a specific form of communication architecture is crucial, as it directly impacts the effectiveness of the protocol in mitigating collisions. Factors such as network size, the density of readers, available resources, and system requirements play a significant role in determining the most suitable communication architecture.

By carefully considering the constraints and requirements of the environment, readers can adopt either a distributed or centralized communication approach to coordinate and ensure efficient operation while minimizing collisions.

2.1. Distributed Algorithms

Pulse [19] is a distributed anti-collision protocol that utilizes a control channel for exchanging notifications between readers. When it comes to tag interrogation, Pulse employs a single data channel. The reader currently engaged in tag reading broadcasts a beacon to prevent neighboring readers from utilizing the active channel. However, in dense RFID networks, this approach leads to the disabling of many readers, resulting in a degradation of system performance.

To address this problem, the MCMAC (Multi-Channel MAC) [20] protocol is introduced. MCMAC allows RFID readers to manage multiple data channels at the MAC layer. Readers use the control channel to announce the occupied frequencies to their neighboring readers, enabling more efficient utilization of resources.

Another distributed TDMA-based protocol, known as Distributed Efficient & Fair Anti-Collision (DEFAR) [21], aims to select at least one reader within a collision domain. It achieves this by exchanging beacons between readers, effectively reducing both mono-channel and multi-channel collisions. Priorities are assigned to readers based on their behavior compared to previous states, enabling the selection of a reader based on its identification and priority level.

For mobile dense and time-critical deployments in RFID networks, Coverage Oriented Reader Anti-Collision (CORA) [22] provides a reciprocal local solution. Each reader depends on its neighborhood for activation, allowing for identification of colliding neighbors. CORA relies on local learning within each reader’s neighborhood.

HAMAC (High Adaptive MAC) [23] is a dedicated protocol designed for mobile and large-scale RFID networks. Based on CSMA (Carrier Sense Multiple Access), HAMAC dynamically controls window contention to avoid collisions in the WSN (Wireless Sensor Network). It adjusts its contention window based on network congestion on available frequencies, employing either linear or multiplicative decreasing mechanisms.

Frequency Time Scheme MAC (FTSMAC) [24] is a distributed RFID anti-collision protocol that employs a special notification system for distributing frequency resources. The first reader to wake up initiates the process of discovering neighboring readers for frequency reuse, and this discovery process continues. Readers within the same data channel group query tags simultaneously during assigned time slots.

DMLAR (Distributed Machine Learning-Based Anti-Collision for Readers) [25] is an innovative anti-collision algorithm designed for RFID readers operating in mobile environments. It offers individual resource control using learned collision prediction models for both Reader-to-Reader and Reader-to-Tag interferences. The dataset continuously updates with each reader movement, and upon simulation completion, readers broadcast their datasets to establish a comprehensive database. The primary objective is to enable collision-free tag interrogation, autonomously improving the efficiency of mobile RFID networks.

DPCCPSO (Distributed Parallel Cooperative Co-evolution Particle Swarm Optimization) [26] presents an enhanced RFID reader anticollision model, which addresses challenges by modifying the measure index, introducing a constraint function, and considering collisions among readers and between readers and tags. The dense deployment of numerous readers leads to a high-dimensional problem, unsolvable by traditional algorithms. Hence, distributed parallel cooperative co-evolution particle swarm optimization (DPCCPSO) is introduced, dynamically adjusting the inertia weight and learning factors during evolution and presenting an improved grouping strategy.

2.2. Centralized Algorithms

A family of centralized anti-collision protocols includes the Neighbor Friendly Reader Anti-collision (NFRA) series, consisting of NFRA, NFRA-C, NFRA-AIC, and Xia-NFRA.

NFRA (Nonlinear Frequency Response Analysis) [27] is a TDMA-based anti-collision protocol designed for dense RFID networks. It utilizes a polling server as a command distributor to schedule reader communications, authorizing a single reader to read tags at a specific time. The NFRA algorithm assumes the presence of a centrally deployed query server, ensuring its messages reach the entire deployment domain. This server control allows readers to operate concurrently without collisions. NFRA adopts a dual-channel strategy, where readers only receive the command signal from the query server and do not respond.

NFRA-AIC (Nonlinear Frequency Response Analysis Adaptive Interrogation Capacity) [28] is an anti-collision algorithm that incorporates adaptive read functionality based on the NFRA system. The key feature is that the query time of a reader is determined based on the number of tags within its range. NFRA-AIC utilizes subrounds to supplement and allocate a time slot after the Access (AC) frame to the first Operation Complete (OC) frame, enabling readers that finish reading tags to release the assigned resource.

NFRA-C (Nonlinear Frequency Response Analysis with Counters) [29] extends the NFRA system by introducing a timer to capture the track record of successful Collision Resolution Tag (CRT) interrogations made by the reader. At each OC frame, the reader broadcasts timers to detect the presence of neighboring readers when encountering a beacon collision. By comparing the timers of colliding readers, the reader with the minimum timer value is granted the CRT interrogation in the current turn.

Xia-NFRA (Xia. Nonlinear Frequency Response Analysis) [30] is an anti-collision algorithm designed specifically for densely populated RFID networks with mobile readers. In this approach, the server waits for a short period to discover the proximity of any reader. During this period, the reader broadcasts a beacon message according to its randomly selected number of periods (TS) and registers the total number of received single-beacon and multiple-beacon messages.

2.3. Protocol Comparisons

In this context, protocols (

Table 1. Protocol comparison.

Attributes	PULSE	DEFAR	MCMAC	CORA	HAMAC	FTSMAC	DMLAR	NFRA	Proposed DTS-AC
RRI	✓	✓	✓	✓	✓	✓	✓	✓	✓
RTI	✓	✓	-	✓	✓	✓	✓	✓	✓
Distributed	✓	✓	✓	✓	✓	✓	✓	-	✓
Centralized	-	-	-	-	-	-	-	✓	-
ML-based	-	-	-	-	-	-	✓	-	-
Multi-data channel	-	✓	✓	✓	✓	✓	✓	-	✓
CSMA	✓	✓	✓	-	✓	✓	-	✓	✓
FDMA	-	-	✓	-	-	✓	✓	-	✓
TDMA	-	✓	✓	✓	-	✓	✓	✓	✓

) are classified based on deployment and attributes such as RRI and RTI resolution, data channels, ML integration, and channel access methods. Our proposed solution, tailored for stable distributed networks, manages multiple data channels to enhance successful communications through a hybrid approach integrating FDMA, TDMA, and channel access control methods. Integrated within this framework is our Distributed Time Slot Anti-Collision protocol (DTS-AC), which employs TDMA for RRI mitigation and FDMA for RTI collision resolution, aligning with our aim of improving communication efficiency in distributed RFID networks.

3. Algorithm Description

In this section, we introduce our novel DTS-AC resource distribution strategy, aimed at improving performance, achieving a stable system, and ensuring efficient allocation of frequency and time resources. The primary goal of this strategy is to address the Reader-to-Reader Interference (RRI) and Reader-to-Tag Interference (RTI) collision problems.

This protocol utilizes a novel TDMA notification system to address RTI concerns while also incorporating the frequency resource management aspect of the FTSMAC (FDMA) [24] protocol to resolve RRI collision issues.

3.1. Background

A.

Control Channel

Our DTS-AC protocol incorporates a notification system to facilitate communication between readers through a designated control channel (Figure 2). We utilize the two distinct control channels to efficiently manage resource allocation.

Control Channel Structure:

• Control Channel 1 (CC1): Dedicated to distributing frequencies for reuse.
• Control Channel 2 (CC2): Utilized for managing TS, enabling effective avoidance of RTI collisions.

To mitigate RRI collisions, a reader identifies the presence of neighboring readers within its interference field on the data channel. It then employs the CC2 control channel to distribute TSs among the neighboring readers, thus preventing RTI collisions. Additionally, the CC2 control channel is utilized to distribute frequencies to neighboring readers, effectively preventing RRI collisions.

B.

Frame Structures and Tables

To control the assignment of frequencies and time slots, our protocol employs specific frame structures and tables within the notification system.

B.1 Time Slot Control

The control messages used (

Table 2. Proposed structure of control messages for Time Slot management.

Message Type	READER_ SENDER	READER_ RECEIVER	ELIMINATE_TS
REQUEST1	✓	_	✓
REQUEST2	✓	_	_
RESPONSE1	✓	✓	_
RESPONSE2	✓	✓	_
TS_select	_	✓	_

) are as follows:

• TYPE: Message type (e.g., REQUEST1, REQUEST2, RESPONSE1, RESPONSE2, TS_select).
• READER_SENDER: Identification of the source reader.
• READER_RECEIVER: Identification of the destination reader.
• ELIMINATE_TS: TS used by the source reader.

The information stored in the readers’ table is as follows:

• TS_Enable: Authorized Time Slots for use.
• TS_Disable: Time Slots not allowed for use.

B.2 Frequency Control

The control messages used (

Table 3. Proposed structure of control messages for frequency management.

Message Type	READER_SENDER	READER_ RECEIVER	READER_IN_CHAIN	AFECT _FREQ
REQUEST1	✓	-	✓	-
REQUEST2	✓	-	-	-
RESPONSE	✓	✓	-	-
ADD_TO_CHAIN	-	✓	-	✓
NEW_CHAIN	-	✓	-	✓

) are as follows:

• USED_PROTOCOL: Protocol used, either FTDMA or CSMA.
• READER_IN_CHAIN: Readers using the same data channel.
• AFFECT_FREQ: Affected frequency.

The information stored in the readers’ table is as follows:

• TYPE: Message type (e.g., REQUEST1, REQUEST2, RESPONSE, ADD_TO_CHAIN, NEW_CHAIN).
• READER_SENDER: Sender reader ID.
• READER_RECEIVER: Receiver reader ID.
• READER_IN_CHAIN: Readers using the same data channel.
• AFFECT_FREQ: Assigned frequency.

3.2. Algorithm

A.

Time Slot Control Process

The objective of the algorithm depicted in Figure 3 and Figure 4 is to establish a distributed system that effectively prevents readers from utilizing TSs that could potentially result in RTI.

• Initially, all readers are in the Backoff state [24]. When a reader initiates the process, it begins the notification procedure with its neighboring readers.
• During this phase, the reader enters the receive phase (green zone) upon receiving a control message from another reader. If no control message is received, it proceeds with the send and read phase (blue zone).

A.1

Sending and Reading phase

• The reader initiates tag interrogation using the data channel if there is an available TS in its table (IsEmpty(TSEnable) = 0). Otherwise, it selects an available TS from the list and stores it in TSEnable.
• Subsequently, the reader broadcasts a REQUEST1 message through the control channel to identify the nearest neighbor potentially affected by RTI. This serves as an alert to neighbors to avoid using the reader’s TS.
• After a delay period of Tmin, if the reader receives a RESPONSE1 message from the targeted neighbors, it retains the first received message. The reader then activates the sender of the RESPONSE1 message through a TS_select message, enabling them to select their TS and initiate the subsequent discovery process.
• If no response is received within Tmin, the reader broadcasts a new REQUEST2 message to discover other readers, utilizing existing neighbors as relays.
• In the case of receiving a RESPONSE1 message, the reader retains the first message and sends a TS_select message to the new neighbor, initiating the next phase of the process.

A.2

Reception phase

When a reader receives a REQUEST1 or REQUEST2 message on the control channel, its response varies based on its current state.

Response to REQUEST1:

• If the reader receives a REQUEST1 message, it updates its table entry by adding the TS value from the ELIMINATE_TS field of the corresponding RESPONSE1(TS) frame to the TS_Disable column. This is achieved by the equation TSDisable = TSDisable + RESPONSE1(TS).
• Subsequently, if there are available Time Slots (IsEmpty(TSEnable) = 0) in its table, the reader sends a RESPONSE1 message. It selects one of the available TSs and stores it in the TS_Enable field (TSEnable = TS). If there are no available TSs, the reader does not participate in the process.

Response to REQUEST2:

• If the reader already has a TS and is unaffected (IsEmpty(TSEnable) = 0), it acts as a forwarding node. The reader rebroadcasts the REQUEST2 frame to its neighboring readers.
• Upon receiving a RESPONSE2 message, the reader forwards it to the sender of the original REQUEST2 message and waits for a response containing the TS value to be forwarded.
• However, if the reader is affected because it does not have a TS (IsEmpty(TSEnable) = 1), it sends a RESPONSE2 message and waits for a TS_select response. This response provides the necessary TS for the reader to participate in the process.

By following these actions, the readers effectively respond to different control channel messages, either updating their table entries or forwarding messages to ensure a coordinated and efficient allocation of TS.

B.

Frequency Control Process

In the frequency control part (algorithm in the Appendix A and Figure 5), readers begin by waiting for their backoff period to ensure that no other reader initiates the frequency assignment process.

B.1

Sending and Reading phase

Following the backoff period, the reader proceeds with frequency control. If CSMA is indicated, it awaits beacon messages on the control channel for a specific period (Tmin). In the absence of beacons, the reader designates a free frequency in its table as “FDMA”. Subsequently, the reader initiates communication with neighboring readers through REQUEST1 frames, with the aim of frequency reuse. In the event of no response, the reader broadcasts a REQUEST2 message to identify a reader using a new frequency.

Upon completing the sending phase, the reader can utilize the data channel for tag queries. If the reader has an assigned frequency (FDMA), it accesses the concurrent access channel directly. Otherwise, if no frequency is assigned, the reader resorts to CSMA for accessing the concurrent access channel.

By following these systematic steps, the reader efficiently establishes frequency allocation, engages in communication with neighboring readers, and accesses the data channel.

B.2

Reception phase

During the backoff phase of the frequency control process, a reader may receive requests from other readers. If a REQUEST1 frame is received, the reader evaluates whether it is affected by the request based on a predefined threshold probability (Pr < Threshold). Additionally, the reader checks for any interference from readers listed in the READER_IN_CHAIN list.

If no interference is detected, the reader transitions from Carrier Sense Multiple Access (CSMA) to Frequency Time Division Multiple Access (FTDMA). It then adds its identification to the READER_IN_CHAIN list and listens for the ADD_TO_CHAIN message to utilize the affected frequency. However, if the reader is unaffected by the request or if interference is present, it retains the current READER_IN_CHAIN list in its table.

In the event of receiving a REQUEST2 message, the reader responds with a RESPONSE frame. If a NEW_CHAIN message is received, indicating the need for a new communication chain, the reader switches from CSMA to FTDMA. Additionally, it selects a new frequency and adjusts its operation accordingly. If no NEW_CHAIN message is received, the reader returns to the IDLE state.

These actions ensure that readers in the frequency control process respond appropriately to requests and interference, facilitating efficient frequency assignment and allowing for adaptability to changing network conditions.

4. Simulation and Results

In this section, the simulation results of an RFID system applied to a wireless sensor network are presented. The system consists of randomly distributed mobile readers and RFID tags. The objective is to investigate the collision environment in the RFID system. The key simulation parameters are provided in

Table 4. Simulation parameters.

Parameter	Value
Number of readers (sparse network)	10, 20, …, 50
Number of readers (dense network)	50, 100, …, 250
Number of tags	100
Reader and Tag position	Random
Type of antenna	Omni-directional
Read range of data channel (rr)	3, 5 m
Read range of control channel (crr)	2 × rr
Number of Data Channel	7
Number of Time Slot (case 1)	7
Number of Time Slot (case 2)	1, 2, 3, 4, 5, 6, 7
Number of control channel	1
protocols compared	FTSMAC, Pulse, MCMAC, CORA
Backoff	(ReaderID − 1) × CW
CW	Convergence time of all readers
Tmin	5 ms (Standard EPC)
T	Neighboring readers’ response time

.

The performance of five distributed collision avoidance protocols from the literature—FTSMAC [24], Pulse [19], MCMAC [20], CORA [22] and NFRA [27]—is compared based on the obtained results.

• Pulse: Utilizes a control channel for notifications between readers and a single data channel for tag interrogation.
• MCMAC: Allows RFID readers to manage multiple data channels, announcing occupied frequencies via a control channel for more efficient resource utilization.
• CORA: Designed for mobile, dense deployments in RFID networks; CORA relies on local learning within each reader’s neighborhood to activate and identify colliding neighbors.
• FTSMAC: Utilizes a distributed RFID anti-collision protocol with a notification system for distributing frequency resources, enabling readers to query tags simultaneously in assigned time slots within the same data channel group.
• NFRA: Based TDMA protocol for dense RFID networks, coordinating reader communications via a polling server and ensuring collision-free operation through centrally deployed query servers.

These protocols are evaluated in terms of their ability to mitigate collisions effectively.

The simulation is implemented using MATLAB 2019a and comprises various modules, including Wireless Sensor Network Mobility, the Reader RFID Module, the Tag RFID Module, Collision RRI, Collision RTI, Reader–Tag Communication, Reader–Reader Communication, and the Anti-Collision Protocol.

By analyzing the simulation results, insights are gained into the collision behavior of RFID systems. The performance of the different protocols is assessed, providing a better understanding of their effectiveness in collision avoidance and optimizing communication between readers and tags in the wireless sensor network.

4.1. Interference Modeling

In essence, saying a method can handle RRI and RTI collisions does not guarantee they will not happen, but rather that the method has ways to lessen their impact. Proof involves showing how the method’s design or algorithm manages resources and scheduling to minimize collisions, typically through mathematical modeling, simulation, or performance analysis under different conditions.

In our study, we introduce a novel formula to quantify the number of active readers within a TDMA-based RFID system, considering multiple simulation periods (Ps) and available time slots (TSs).

A.

Number of Active Readers:

$$\begin{gathered} AR={\sum }_{i=1}^P{\sum }_{j=1}^{TS}\frac{TDMA\_Schem(i,j)}{P} \\ \left\{\begin{array}{c} AR Maximize \\ TS Minimise\end{array}\right. \end{gathered}$$

(1)

The following formula illustrates the dynamic interaction between simulation periods, time slots, and the reuse of time slots by multiple readers. Here, TDMA_Schem(i,j) denotes the number of readers simultaneously accessing a particular time slot during the ith simulation period, with P representing the total number of simulation periods and TS representing the available time slots.

Furthermore, our objective functions, stipulating the maximization of active readers while potentially minimizing the number of time slots TS, highlight the trade-off between system performance and resource allocation. This optimization framework enables the design and implementation of TDMA schemes tailored to specific operational requirements, balancing throughput and resource utilization in RFID communication systems.

B.

Number of Collisions:

Collision events, such as Reader-to-Reader Interference (RRI) and Reader-to-Tag Interference (RTI), significantly impact the performance and reliability of Radio Frequency Identification (RFID) systems. In our study, we propose a comprehensive formula to quantify the total number of collision occurrences within an RFID network across multiple simulation periods.

The formula, denoted as

$$\begin{gathered} C={\sum }_{i=1}^P{\sum }_{j=1}^{R}RRI(i,j,f)+RTI\left(i,j,f,ts\right) \\ RRI,RTI=\left\{\begin{array}{c} 0 No Collision\\ 1 Collision\end{array}\right. \\ \left\{\begin{array}{c} RRI Minimise\\ RTI Minimise\end{array}\right. \end{gathered}$$

(2)

captures the cumulative impact of RRI and RTI events over P simulation periods. Here, R represents the total number of readers in the system, and P signifies the number of simulation periods.

The component RRI(i,j,f) indicates whether the jth reader at the ith period enters into an RRI-type collision using frequency f. Similarly, RTI(i,j,f,ts) indicates the case of reader interference entering an RTI-type collision using frequency f and time slot ts.

The binary representation of collision events (0 for no collision, 1 for collision) facilitates the minimization of collision occurrences for both RRI and RTI scenarios. By minimizing collision events, the formula aids in optimizing system throughput and reliability, enhancing the overall performance of RFID deployments.

4.2. Analysis of Results

In this section, we analyze the performance of our protocol using various experiments and scenarios to demonstrate its effectiveness in minimizing collisions.

Figure 6 illustrates the significant improvement of DTS-AC compared to FTSMAC across various numbers of reader deployments. Since DTS-AC minimizes the RTI collisions affecting the FTSMAC protocol, its schema-based collision management consumes more Time Slots. As the number of readers increases, the system’s reading efficiency experiences a significant improvement, with efficiency reaching 88% when 30 readers are present. Subsequently, as the number of readers continues to increase, the reading efficiency approaches almost 100%, indicating a near-perfect performance. The frequency and time resource distribution technique employed by both DTS-AC and FTSMAC demonstrates exceptional system performance, surpassing that of Pulse, MCMAC, and CORA protocols.

Figure 7 confirms the superior performance of DTS-AC regarding the number of active readers, demonstrating that almost all readers in the simulation can be activated and utilize the data channel without collisions. Both protocols exhibit symmetrical evolution, with DTS-AC showing a slight advantage of three to nine drives. However, a notable deviation between DTS-AC and FTSMAC becomes apparent at 50 readers, indicating DTS-AC’s more effective collision minimization while maintaining robust RRI processing and enhancing the RTI resolution process.

In Figure 8, a study of a dense RFID network is conducted with varying numbers of readers (ranging from 50 to 250) across seven data channels and time slots. The results reveal a substantial disparity between the two protocols compared to the earlier sparse network scenario. In this denser deployment, DTS-AC has evolved significantly, facilitating the activation of up to 230 readers. The significant gap of 160 active readers between the two protocols underscores the effectiveness of DTS-AC’s time slot management, highlighting its pivotal role in accommodating such an increase. This underscores DTS-AC’s suitability for dense RFID systems.

Figure 9 illustrates the system’s efficiency across various numbers of readers (ranging from 10 to 50) and the utilization of time slots (from one to seven TSs). DTS-AC consistently outperforms FTSMAC in all scenarios, with performance differentials ranging from 12% to 18%. Remarkably, even with the escalation in reader deployments, DTS-AC sustains its superiority, achieving a performance peak of 98% with 50 readers. This highlights the critical role of time slot control by DTS-AC in optimizing system efficiency amidst varying network densities and demands.

Considering the cost of the solution, Figure 10 shows the number of failed reader interrogations. It is evident that DTS-AC minimizes the delays required for readers to interrogate all tags in the network, with delays ranging up to 5 s for 50 readers. In contrast, FTSMAC exhibits a longer delay of 1.5 s and an increased number of failed interrogations, reaching 120 for 50 readers. NFRA, a centralized algorithm, exhibits longer interrogation times due to control messages exchanged between the reader and server, but with no failed interrogations. MCMAC shows similar delays to FTSMAC initially but exhibits the highest number of failed interrogations, reaching 160 for 50 readers.

These findings demonstrate the superior performance, efficiency, and cost-effectiveness of DTS-AC compared to other protocols in various simulation scenarios.

4.3. Complexity Analysis

The following analysis provides a comprehensive examination of the complexity associated with each element of this protocol (Figure 11), based on the provided parameters and their respective weights. Each element, including Active Reader, Collision, Delay, and Overhead, is dissected to understand how various factors contribute to the overall complexity of the protocol. By examining the dependence of parameters such as the number of readers and frequency, time slots, and canal ranges, this analysis aims to shed light on the intricacies of the protocol, enabling a deeper understanding and potential avenues for optimization. Let us delve into the detailed analysis of each element:

To create a complexity formula for each element based on its parameters, we use a weighted sum approach, where each parameter is multiplied by its corresponding weight. Let us designate the parameters as follows:

• R: Number of readers;
• T: Number of tags;
• F: Number of frequency;
• S: Number of time slots;
• D: Range of data channel;
• C: Range of control channel.

The complexity formulas are as follows:

• Active Reader (A)

A = 0.4R + 0.2T + 0.7F + 0.6S + 0.8D + 0.9C

The complexity of active reader management in the DTS-AC protocol is heavily influenced by the allocation of frequency and time resources, with significant weight assigned to these parameters. Additionally, the number of readers and the range of control channels play important roles in determining the overall complexity.

2.

Collision (Cn)

Cn = 0.4R + 0.08T + 0.6F + 0.6S + 0.5D + 1C

The collision management complexity is primarily driven by the number of readers and the range of control channels, which are assigned the highest weights. The allocation of frequencies and time slots also contributes significantly to the overall complexity of collision resolution.

3.

Delay (Dl)

Dl = 0.7R + 0.2T + 0.8F + 0.9S + 0.9D + 0.6C

Delay analysis complexity is predominantly influenced by the number of readers and the range of data channels, both of which carry substantial weight. Additionally, the allocation of frequencies and time slots plays a significant role in determining the delay management complexity.

4.

Overhead (O)

O = 0.94R + 0T + 0.6F + 0.6S + 0.56D + 0.9C

The overhead complexity is primarily determined by the number of readers, which carries the highest weight. The range of control channels and the allocation of frequencies also contribute significantly to the overall overhead management complexity.

5. Conclusions and Future Work

In our study, we introduced the Distributed Time Slot Anti-Collision protocol (DTS-AC), a novel MAC layer protocol specially designed to address the challenges inherent in dense RFID networks. DTS-AC is engineered to mitigate the interference issues stemming from RRI and RTI collisions, which can significantly hinder system performance and resource utilization.

DTS-AC incorporates a sophisticated TDMA notification system to manage RTI concerns effectively. This notification system enables efficient coordination among readers, minimizing collisions between reader transmissions and tag responses. Additionally, DTS-AC leverages elements of the FTSMAC protocol, particularly its frequency division multiple access (FDMA) scheme, to manage frequency resources and alleviate RRI collisions. By dynamically allocating time slots and frequencies, DTS-AC optimizes resource utilization, reduces contention, and enhances overall system stability.

Our research findings underscore the superior performance of DTS-AC compared to established protocols like FTSMAC, Pulse, MCMAC, and CORA across diverse simulation scenarios. Notably, DTS-AC demonstrates significant improvements in read efficiency, the activation of readers, and collision resolution, even in densely populated RFID networks. These enhancements translate into tangible benefits such as reduced interrogation delays and minimized failed reader interrogations, thereby enhancing the reliability and efficiency of RFID systems.

By addressing the challenges of interference and contention in dense RFID environments, DTS-AC not only advances the state of the art in RFID technology but also lays the groundwork for further advancements in MAC protocols for multi-channel RFID networks. Future research endeavors could explore additional optimization techniques and practical implementations to further enhance the performance and applicability of DTS-AC in real-world deployments.