In the context of robotics, the implementation of Networked Control Systems (NCS), especially when based on wireless networks, introduces significant flexibility but also challenges related to transmission delays. The non-deterministic nature of such delays can degrade performance and compromise the stability of the control system. This thesis aims to analyze and model, within this framework, the impact of modern IEEE 802.11ax (Wi-Fi 6) networks. The work consists of an experimental campaign designed to stochastically characterize the One-Way Delay (OWD) and the Round-Trip Time (RTT). The measurements were carried out using a testbed composed of two Raspberry Pi 4 Model B boards acting as IEEE 802.11ac (Wi-Fi 5) clients and a TP-Link AX1500 router serving as an IEEE 802.11ax Access Point (AP), replicating a typical mixed configuration. Temporal synchronization between nodes was achieved through a local Network Time Protocol (NTP) server. The tests were conducted using the Robot Operating System 2 (ROS 2) Humble Hawksbill middleware and custom-developed C++ nodes. The collected data were then statistically analyzed, using Python scripts, to derive a probabilistic model of the communication channel. The results show that the delay can be described by a Lognormal random variable whose parameters depend on operational factors such as message payload size. Finally, the stochastic channel model was integrated into a MATLAB Simulink simulation environment to investigate the effects of the measured delays on the control of an ABB IRB120 industrial robotic arm, analyzing the resulting performance degradation.
Nel contesto della robotica, l’implementazione di sistemi di controllo in rete (Networked Control Systems, NCS), specialmente se basati su reti wireless, introduce vantaggi di flessibilità ma anche criticità legate ai ritardi di trasmissione. La natura non deterministica di questi ultimi può degradare le prestazioni e compromettere la stabilità del sistema di controllo. Questa tesi si propone di analizzare e modellare in questo contesto l'impatto delle moderne reti IEEE 802.11ax (Wi-Fi 6). Il lavoro consiste in una campagna sperimentale volta a caratterizzare stocasticamente il ritardo di comunicazione unidirezionale, o One-Way Delay (OWD), e il tempo di andata e ritorno, o Round-Trip Time (RTT). Per condurre le misurazioni, è stato allestito un testbed composto da due Raspberry Pi 4 Model B, client 802.11ac (Wi-Fi 5), e un router TP-Link AX1500, Access Point (AP) 802.11ax, replicando un comune scenario misto. La sincronizzazione temporale tra i nodi è stata ottenuta tramite un server Network Time Protocol (NTP) locale. I test sono stato condotti utilizzando il middleware Robot Operating System 2 (ROS2) Humble Hawksbill e nodi C++ sviluppati ad hoc. Successivamente, i dati raccolti sono stati analizzati statisticamente, tramite script Python, per sintetizzare un modello probabilistico del canale. Si è dimostrato come il ritardo possa essere descritto da una variabile aleatoria Lognormale i cui parametri dipendono da fattori operativi, come la dimensione del payload dei messaggi. Infine, il modello stocastico del canale è stato integrato in un ambiente di simulazione MATLAB Simulink per investigare gli effetti dei ritardi così ottenuti sul controllo di un braccio robotico industriale ABB IRB120, analizzando il degrado delle prestazioni.
Impatto delle reti wireless sul controllo di robot
MARTINI, NICOLÒ
2024/2025
Abstract
In the context of robotics, the implementation of Networked Control Systems (NCS), especially when based on wireless networks, introduces significant flexibility but also challenges related to transmission delays. The non-deterministic nature of such delays can degrade performance and compromise the stability of the control system. This thesis aims to analyze and model, within this framework, the impact of modern IEEE 802.11ax (Wi-Fi 6) networks. The work consists of an experimental campaign designed to stochastically characterize the One-Way Delay (OWD) and the Round-Trip Time (RTT). The measurements were carried out using a testbed composed of two Raspberry Pi 4 Model B boards acting as IEEE 802.11ac (Wi-Fi 5) clients and a TP-Link AX1500 router serving as an IEEE 802.11ax Access Point (AP), replicating a typical mixed configuration. Temporal synchronization between nodes was achieved through a local Network Time Protocol (NTP) server. The tests were conducted using the Robot Operating System 2 (ROS 2) Humble Hawksbill middleware and custom-developed C++ nodes. The collected data were then statistically analyzed, using Python scripts, to derive a probabilistic model of the communication channel. The results show that the delay can be described by a Lognormal random variable whose parameters depend on operational factors such as message payload size. Finally, the stochastic channel model was integrated into a MATLAB Simulink simulation environment to investigate the effects of the measured delays on the control of an ABB IRB120 industrial robotic arm, analyzing the resulting performance degradation.| File | Dimensione | Formato | |
|---|---|---|---|
|
Martini_Nicolo.pdf
Accesso riservato
Dimensione
1.42 MB
Formato
Adobe PDF
|
1.42 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/97704