Swarm Robot Communication Using ESP-NOW Mesh Protocol for Multi-Agent Coordination in Indoor Navigation

The development of robotics technology has shown a very significant increase over the past two decades, driven by a growing tendency toward more intelligent, collaborative, and adaptive system designs capable of operating in dynamic environments [1]. One rapidly evolving field within this domain is swarm robotics, a multi-robot approach inspired by the social behaviors of biological organisms such as ants, bees, and fish, which accomplish collective tasks through local interactions and self-organization [2, 3]. This approach offers scalability, redundancy, and high fault tolerance, as the system does not rely on a single centralized controller [4]. When one agent fails, other agents can compensate, allowing the mission to continue collectively and efficiently [5, 6].

On the other hand, the rapid advancement of the Internet of Things (IoT) has enabled physical devices to communicate and exchange data wirelessly in real time for monitoring and adaptive control purposes [7]. Initially applied to static sensor networks, IoT has evolved into IoT-based robotic systems, in which robots act as mobile IoT nodes that perform not only sensing and actuation but also intelligent inter-unit communication [8, 9]. The integration of IoT and robotics has given rise to the paradigm of IoT-based swarm robotics, which enables the emergence of collective intelligence through distributed network architectures [10, 11]. This paradigm forms the foundation of cyber-physical swarm systems capable of performing navigation, environmental mapping, and mission coordination in a decentralized manner [12].

Nevertheless, the implementation of swarm robotics in indoor environments still faces significant challenges, particularly in terms of inter-robot communication. Conventional Wi-Fi technology requires access points and suffers from limitations in the number of simultaneous connections [13]. Bluetooth provides limited bandwidth and short communication range, which are insufficient for high-speed swarm data exchange [14, 15]. Meanwhile, LoRa excels in long-range communication but is unsuitable for real-time multi-robot coordination due to its very low throughput capacity [16]. These limitations highlight the need for communication protocols that operate without infrastructure, offer low latency, consume minimal energy, and support a large number of nodes for indoor swarm operations [17, 18].

One communication protocol that has recently attracted increasing attention is Enhanced Peer-to-Peer Network (ESP-NOW), a low-power, low-latency communication protocol based on ESP32 that enables peer-to-peer data exchange without routers or internet connectivity [19]. Its advantages, including small packet size, low energy consumption, and multi-node communication support, make ESP-NOW highly promising for swarm robotics applications [20, 21]. Unlike conventional swarm communication frameworks that rely on centralized network infrastructure, the proposed ESP-NOW Mesh model emphasizes full decentralization with peer-to-peer synchronization. This design choice positions the proposed system as one of the first implementations focusing on infrastructure-free real-time swarm coordination in indoor environments. However, most existing ESP-NOW-based studies still focus on static IoT sensors, environmental monitoring, or single-node usage rather than dynamic communication among mobile robots [22]. Research on the application of ESP-NOW in multi-agent mesh schemes for indoor navigation remains very limited, particularly with respect to latency evaluation, packet delivery ratio (PDR), Received Signal Strength Indicator (RSSI) degradation, and formation accuracy [23, 24].

Furthermore, many studies on indoor swarm robotics reported in the literature remain dominated by simulations and lack real-hardware implementations operating under actual conditions [25]. In practice, signal interference, multipath fading, room geometry, and robot motion dynamics represent critical factors that cannot be fully captured through digital simulations alone [26, 27]. Therefore, experimental evaluations based on real implementations are far more relevant for understanding network stability, coordination reliability, and navigation performance under realistic conditions [28]. This real-hardware experimental emphasis distinguishes the proposed study from simulation-dominated swarm research and strengthens the empirical validity of the obtained performance results.

Based on these research gaps, this study proposes the development and evaluation of an IoT-based swarm robot system using the ESP-NOW Mesh communication protocol for multi-agent coordination in indoor navigation. Each robot functions as a mobile IoT node that exchanges real-time information related to position, movement direction, and sensor status. Experimental evaluations are conducted to measure PDR, communication latency, RSSI, and formation success across various indoor scenarios [29, 30]. With this perspective, the proposed framework contributes not only as a communication solution but also as a benchmark architecture for future IoT-based swarm robotic research. Consequently, this work provides not only technical contributions to the implementation of IoT-based swarm robots but also opens new research directions toward more scalable distributed multi-robot systems suitable for real operational environments [31-33].

This research is expected to serve as a foundation for the development of large-scale IoT-based swarm robotics systems for indoor monitoring, warehouse automation, building exploration, and disaster victim search operations [34]. In addition to providing a lightweight and efficient communication framework, the results of this study also create opportunities for the development of more adaptive, intelligent, and artificial intelligence-oriented coordination algorithms in future research.

Beyond experimental validation, the results of this study also provide practical design guidelines for implementing decentralized indoor swarm robotic systems based on ESP-NOW Mesh communication. The quantified performance boundaries obtained—such as communication reliability degradation, latency tolerance thresholds, and effective communication range—can be used as reference parameters for designing scalable and reliable IoT-based swarm platforms. Thus, the proposed framework can serve as an engineering benchmark for the development and deployment of indoor swarm systems in various scenarios, particularly warehouse automation, collaborative inspection, and multi-robot building exploration.

The main contributions of this research can be summarized as follows:

1. Development of a decentralized swarm robot communication framework based on the ESP-NOW Mesh protocol for indoor multi-agent coordination.
2. Experimental evaluation of the communication performance using four key metrics: PDR, communication latency, RSSI signal strength, and Formation Stability Index (FSI).
3. Validation of swarm formation stability under multiple indoor navigation scenarios, including straight corridor navigation, obstacle avoidance, and multi-room path splitting.
4. Introduction of the FSI as a practical metric for evaluating formation coordination in real swarm robotic systems.

Table 1. Comparison of communication technologies for multi-robot systems

Although ESP-NOW has been widely applied in IoT-based communication systems, most existing implementations focus on static sensor networks or single-node communication scenarios. The application of ESP-NOW in dynamic multi-robot systems, particularly for real-time swarm coordination in indoor environments, remains relatively unexplored. Therefore, this study aims to investigate the feasibility and performance of ESP-NOW Mesh communication for real swarm robotic platforms operating under indoor navigation conditions. A comparative summary of widely adopted communication technologies for multi-robot systems is presented in Table 1, highlighting their respective characteristics in terms of latency, communication range, power consumption, and suitability for swarm robotic applications.

This section presents the experimental evaluation of the proposed IoT-based swarm robot communication system using the ESP-NOW Mesh protocol in an indoor environment. The experiments were conducted under three navigation scenarios: straight corridor (S1), obstacle-centered path (S2), and split-path environment (S3). The performance of the system was assessed using four key metrics: PDR, communication latency, RSSI, and FSI. To ensure statistical reliability, each experiment was repeated five to eight times, and the reported values represent the average results obtained from these trials.

Table 8. Evaluation parameters

Table 8 presents the key performance metrics used to evaluate the proposed swarm robotic system. These parameters collectively measure communication reliability, network responsiveness, signal propagation characteristics, and swarm coordination stability.

3.1 Packet Delivery Ratio performance

PDR represents the success rate of packet transmission between robots in the swarm. As the proposed system employs peer-to-peer broadcast communication, an increase in the number of nodes leads to higher channel contention and network congestion, which may reduce the number of successfully received packets.

Table 9 presents the average PDR values obtained under three indoor navigation scenarios with varying swarm sizes. For the configuration with three robots, PDR values remain above 98% across all scenarios, indicating that the ESP-NOW Mesh communication provides very high reliability for small-scale swarms. When the swarm size increases to five robots, the PDR slightly decreases but remains above 95%, demonstrating that the system is still able to maintain stable swarm synchronization despite increased broadcast traffic density. For the seven-robot configuration, a more pronounced PDR reduction is observed, particularly in Scenario S3 (split path), which is influenced by increased packet collisions and RSSI degradation caused by multi-hop routing and multipath fading in branched indoor environments. Nevertheless, PDR values above 92% are still considered acceptable for real-time swarm coordination, confirming that the ESP-NOW Mesh protocol remains reliable up to a medium-scale swarm size.

Table 9. Average Packet Delivery Ratio (PDR) results (% over 5–8 trials)

Figure 6. Packet Delivery Ratio (PDR) trend versus number of swarm nodes

Figure 6 illustrates the trend of PDR as the number of robots in the swarm increases. For configurations with three and five robots, PDR values remain consistently high, exceeding 95%, indicating that the ESP-NOW Mesh network maintains reliable communication and stable broadcast synchronization at small to medium swarm scales. Under these conditions, channel contention and packet collisions remain relatively manageable, allowing collective swarm decisions to be executed consistently. As the number of robots increases to seven nodes, a more noticeable decline in PDR is observed, primarily due to increased channel contention, a higher number of broadcast collisions, and multipath fading effects in complex indoor environments. Despite this degradation, the PDR remains above 92%, which is still sufficient to support real-time swarm coordination.

3.2 Latency analysis in multi-agent communication

Latency is measured as the time difference between packet transmission from the leader robot and reception by follower robots within a single hop. Low latency is a critical requirement for swarm coordination, as propagation delays may cause followers to react late to formation correction commands.

$Latency$ $=T_{\text {receive}}-T_{\text {send}}$

Table 10. Average communication latency

Table 10 summarizes the average communication latency measured under three indoor navigation scenarios with different swarm sizes. For the configuration with three robots, latency values range from 12.6 to 16.8 ms, indicating very fast inter-node data exchange that effectively supports real-time swarm coordination. When the swarm size increases to five robots, latency rises to the range of 18.9–25.7 ms, reflecting increased channel contention and the need for multi-hop routing in more complex navigation scenarios. For the seven-robot configuration, latency increases further to 27.3–36.8 ms, particularly in Scenario S3 (split path), due to a higher volume of broadcast packets, intensified multi-hop relaying, and multipath fading effects in indoor environments. Nevertheless, latency values remaining below 40 ms indicate that the system is still capable of maintaining stable coordination and formation correction at a medium swarm scale.

Figure 7. Latency comparison across indoor navigation scenarios

Figure 7 compares communication latency across the three indoor navigation scenarios, namely Straight Corridor (S1), Obstacle Center (S2), and Split Path (S3), for different swarm sizes. Scenario S1 consistently yields the lowest latency because the straight navigation path minimizes multi-hop routing requirements and reduces packet collision probability. In Scenario S2, latency increases due to the presence of physical obstacles that trigger trajectory changes and additional broadcast communication during collective avoidance. Scenario S3 exhibits the highest latency across all swarm configurations, as branched paths significantly increase multi-hop routing requirements and amplify channel contention and multipath fading. These latency variations demonstrate that indoor spatial complexity has a substantial impact on mesh communication performance and directly influences the stability of swarm coordination during navigation.

3.3 Received Signal Strength Indicator degradation over indoor distance

RSSI measurements were conducted to evaluate the degradation of ESP-NOW signal strength as a function of inter-robot distance in indoor environments. The RSSI behavior is modeled using the logarithmic path-loss model:

$R S S I(d)=R S S I_0-\left(10 n \log _{10} d\right)$

Table 11. Received Signal Strength Indicator (RSSI) versus indoor distance

Table 11 presents the relationship between inter-robot distance and the RSSI in an indoor environment. The results show a consistent degradation of RSSI as distance increases, from −38 dBm at 1 meter to −81 dBm at 10 meters. This decreasing trend reflects the typical characteristics of indoor wireless signal propagation, which is strongly influenced by multipath fading, wall reflections, and the presence of physical obstacles. An RSSI level of approximately −80 dBm at a distance of 10 meters indicates that ESP-NOW communication can still be maintained near the effective indoor range limit, albeit with a reduced signal margin. This RSSI degradation correlates with the observed increase in latency and reduction in PDR for larger swarm configurations, thereby highlighting the importance of multi-hop routing mechanisms in preserving mesh communication continuity among robots. These results suggest that the practical communication range of ESP-NOW in indoor swarm environments lies within approximately 8–10 meters, making multi-hop routing a critical requirement for reliable swarm connectivity.

Figure 8 illustrates the RSSI degradation curve as a function of inter-robot distance in an indoor environment. The signal strength decreases in a logarithmic manner as distance increases, which is a common characteristic of radio wave propagation in enclosed spaces affected by multipath fading and wall reflections. At distances below 5 meters, RSSI degradation remains relatively moderate, allowing peer-to-peer communication to operate with a sufficient signal margin. However, as the distance increases to 8–10 meters, RSSI values decline sharply toward the receiver sensitivity threshold of the ESP-NOW module. Under these conditions, the probability of packet loss increases, and multi-hop routing becomes necessary to maintain mesh communication reliability. This curve confirms that inter-robot spacing and swarm topology are critical factors in sustaining network stability and multi-robot coordination performance in indoor environments.

Figure 8. Received Signal Strength Indicator (RSSI) degradation curve as a function of distance

3.4 Formation Stability Index

The FSI is used to quantify the ability of the swarm to maintain the desired inter-robot distance during navigation. FSI provides a normalized measure of formation quality by capturing the cumulative deviation of relative distances among swarm members.

$F S I=1-\frac{\sum_{i=1}^N\left|E_D\right|}{N \cdot D_a}$

where, $E_D$ represents the distance error between the actual and desired inter-robot distance, $N$ is the number of robots in the swarm, and $D_a$ is the desired formation distance. FSI values above 0.85 indicate highly stable formations, while values between 0.70 and 0.85 represent moderate stability.

Table 12. Formation error and stability index

Table 12 presents the FSI and average formation error for different swarm sizes. As the number of robots increases, the average formation error rises from 0.9 cm for the three-robot configuration to 3.2 cm for the seven-robot configuration. This increase correlates with higher communication latency and RSSI degradation, which affect synchronization accuracy among robots. Despite this trend, the FSI values for all configurations remain within the stable to highly stable range, with values of 0.955 for three robots, 0.910 for five robots, and 0.840 for seven robots. These results demonstrate that the ESP-NOW Mesh-based swarm system is capable of maintaining acceptable formation stability at small to medium swarm scales and confirm the effectiveness of the proposed formation correction mechanism for indoor navigation.

Figure 9. Formation Stability Index (FSI) versus number of swarm nodes

Figure 9 illustrates the trend of FSI as a function of the number of robots in the swarm. The FSI decreases gradually from 0.955 for the three-robot configuration to 0.840 for the seven-robot configuration. This decline reflects the increasing coordination complexity and accumulation of relative position errors caused by RSSI degradation, higher communication latency, and increased packet collisions within the mesh network. Nevertheless, all FSI values remain within the stable to highly stable categories, indicating that the system consistently preserves swarm formation integrity at small to medium scales. This trend confirms that the ESP-NOW Mesh-based formation control strategy effectively maintains coordinated multi-robot behavior in dynamic indoor environments.

3.5 Discussion of swarm communication performance

Based on the overall experimental results, it can be concluded that ESP-NOW Mesh provides a reliable communication backbone for IoT-based swarm robotic systems operating in indoor environments. The system demonstrates high stability for swarm configurations consisting of three to five robots, while the seven-robot configuration begins to exhibit increased network load, reflected in reduced PDR and increased communication latency. Nevertheless, these degradations remain within acceptable tolerance limits for real-time swarm coordination. The consistently low latency values enable responsive multi-agent formation control, ensuring timely correction of relative position and orientation among robots. Although RSSI degradation becomes more pronounced as inter-robot distance and swarm size increase, the implemented multi-hop routing mechanism effectively maintains communication continuity across the swarm. This capability is particularly important in complex indoor layouts where multipath fading and physical obstructions are unavoidable. Furthermore, the FSI results confirm that the swarm is able to preserve formation integrity throughout navigation tasks, even under increased communication stress. The combined evaluation of PDR, latency, RSSI, and FSI demonstrates that communication performance and formation stability are strongly coupled, and that ESP-NOW Mesh is capable of supporting coordinated swarm behavior in realistic indoor conditions.

3.6 Novelty and contribution analysis

The novelty of this research lies in the experimental implementation of an IoT-based swarm robotic system operating in indoor environments using ESP-NOW Mesh communication. Unlike many previous studies that rely primarily on simulations or experiments involving static IoT nodes, this work presents a fully implemented mobile swarm robot platform evaluated under realistic indoor conditions. This real-world experimental validation provides stronger empirical evidence regarding communication reliability and coordination performance in practical swarm robotic systems. Furthermore, the study performs a comprehensive evaluation using four key performance parameters, namely PDR, communication latency, RSSI signal strength, and FSI. This multi-parameter evaluation enables a more complete understanding of system behavior by simultaneously analyzing communication reliability, network responsiveness, signal propagation characteristics, and swarm coordination stability. The scalability of the proposed system is also demonstrated through experiments involving up to seven swarm robots, which remains relatively uncommon in ESP-NOW-based mobile swarm robotics research. In addition, this study introduces the FSI as a quantitative metric to measure formation integrity and coordination quality in multi-robot systems. The proposed metric provides a normalized and intuitive indicator that can be reused in future swarm robotics studies. Overall, these contributions demonstrate that ESP-NOW Mesh is a lightweight and promising communication framework for infrastructure-free IoT-based swarm robotic systems in indoor environments. Compared with previous swarm communication studies that mainly rely on Wi-Fi or Bluetooth-based networks, the proposed ESP-NOW Mesh approach offers a simpler infrastructure-free architecture with lower latency and energy consumption.