Performance Assessment of MQTT Communication over 4G Networks for Connected E-Bike Systems

Read full text
Authors: A. Figueiredo, Tiago Gándara, José Paulo Oliveira Santos
Cited by: 1
References: 23

Abstract

The growing adoption of connected e-bikes requires reliable, secure, and low-latency communication systems that can support continuous telemetry and remote monitoring. This work presents an experimental performance evaluation of the Message Queuing Telemetry Transport (MQTT) protocol operating over a 4G cellular link using a SIMCom SIM7600 modem. The study quantifies the effects of key parameters, including Quality of Service (QoS) level, Transport Layer Security (TLS), broker location, and time of day, on message cycle time and reliability. Tests were performed using both national and international brokers, with and without encryption, and complemented by serial communication delay estimation. Results show that QoS configuration is the most influential factor, with QoS 0 and 1 achieving average cycle times below 200 milliseconds, while QoS 2 introduces additional overhead due to its multi-step acknowledgment mechanism, with a time increase of approximately 33%. Enabling TLS increased latency by only 11%, confirming the feasibility of secure communication with minimal performance degradation. Broker proximity reduced latency by around 16%, and time-of-day variations had a negligible impact. The serial interface contributed less than 15% of the total delay, confirming its marginal role. Overall, the tested architecture, combining MQTT, 4G connectivity, and modular initialization, proves stable, reliable, and suitable for mobile IoT systems such as connected e-bikes, offering practical design guidelines for balancing performance, security, and efficiency.

Fields of Study & Subjects

Vehicular Ad Hoc Networks (VANETs) IoT Networks and Protocols Network Time Synchronization Technologies

Publication details

Journal / source
Proceedings of the 2026 9th International Conference on Information and Computer Technologies
Publisher
ACM
Pages
484–493
Publication date
11 March 2026