The dataset was acquired by periodically sampling a wireless channel with Wi-Fi frames. The main goal is to track the evolution of the channel quality by acquiring key metrics, such as the arrival of the frame to the destination or the delivery latency. The dataset was obtained through a carefully designed testbed based on two Linux-based personal computers with two TP-Link TL-WDN4800 Wi-Fi, which implements the IEEE 802.11n standard. In total, the four Wi-Fi stations (STAs) were connected to four access points (APs), which are configured to operate in four specific non-overlapping frequency channels (1, 5, 9, and 13) within the 2.4 GHz frequency band. The distance between the STA and the associated AP is about 3 to 4 meters. Ethernet Wi-Fi frames were periodically transmitted with a period of 0.5 seconds, and their payload was set equal to 50 bytes. Small payloads are typical in industrial applications and are helpful for sampling a channel because, being small in size, they perturb less the channel they have to mesure. The frames on the four channels are sent synchronously at the same time. For this reason, the two PCs are synchronized using the Network Time Protocol (NTP). To enhance real-time performance and reduce the jitter in the acquisition of the dataset, the RT-Preempt Linux patch was installed on the two PCs. The primary objective of this experimental setup was to periodically monitor and sample the condition of the channels with minimal additional interference in the surrounding environment. To this end, the device driver of the Wi-Fi stations was customized with several key modifications. This was possible by using the specific open-source "ath9k" device driver and the SDMAC framework described in [1]. The most significant changes to the default parameters of a Wi-Fi network are the following: 1. fix the transmission bit rate at 54 Mbps (i.e., the rate adaptation algorithm was not activated); 2. disable automatic frame retransmissions (i.e., ensure that each frame was transmitted only once); 3. eliminate the standard backoff procedure (i.e., it allows frames to be sent immediately after detecting an idle channel); 4. disable the request-to-send/clear-to-send (RTS/CTS) mechanism (i.e., only data frame and related acknowledgment (ACK) frames are transmitted); 5. downgrade the operational mode from IEEE 802.11n to IEEE 802.11g (i.e, advanced features such as frame aggregation are disabled). Given this configuration, each periodic sample of the channel every 0.5 seconds consists of an exchange of the 50-byte Ethernet frame followed by the associated ACK frame. Each transmitting PC logs in software the transmission time of the data frame (just before transmitting it) and the reception of the related ACK frame, and the difference between the latter and the first timestamp is stored in the dataset. Timestamps were obtained by reading the TSC register of the CPU, and the CPU frequency was used to compute the transmission latency. This was done to improve the precision of the obtained timestamp. In addition, the PC logs the reception of the ACK frame. During the acquisition of the dataset, the spectrum was shared with other interfering nodes that create high dynamics in the condition of the channel, which depends on the usage of the shared wireless medium by students and researchers near our laboratory. The dataset is composed of two directories, namely "train" and "test" which contain files (i.e., dataset segments) where the name is composed of the string "ch" followed by a number that identifies the channel in the 2.4 GHz band, the character "_", an integer number identifying the specific dataset segment, and the file extension ".dat". For instance, the file "ch5_16.dat" identifies a database segment with index 16 acquired in channel 5 of the 2.4 GHz band. By analyzing the file's content, each line represents a packet exchange. For instance, the first ten lines of "ch5_16.dat" are: CH5 1 1684094460000 142.7 ACK 58 540 CH5 2 1684094460500 286.1 ACK 58 540 CH5 3 1684094461000 116.3 ACK 58 540 CH5 4 1684094461500 112.5 ACK 58 540 CH5 5 1684094462000 118.8 ACK 58 540 CH5 6 1684094462500 287.8 ACK 58 540 CH5 7 1684094463000 138.2 ACK 58 540 CH5 8 1684094463500 110.4 ACK 58 540 CH5 9 1684094464000 319.4 NOACK -1 540 CH5 10 1684094464500 112.3 ACK 58 540 Each line has the following fields: In particular: • is the transmission channel in the 2.4 GHz band. It can assume the following values "CH1", "CH5", "CH9", "CH13". • is a progressive number initialized to 1 at the beginning of each file. • is the time when the data packet is sent expressed in Unix epoch in milliseconds. The Unix epoch (or Unix time or POSIX time or Unix timestamp) is the number of seconds that have elapsed since January 1, 1970 (midnight UTC/GMT). • is the transmission acquired just before the queuing of the data packet and the reception of the corresponding ACK frame in microseconds. If the ACK frame is not received, the interrupt raised by the driver to notify the ACK timeout is used. • can assume two values: "ACK" if the ACK frame is correctly received or "NOACK" if an ACK timeout occurred. • is an indication of the signal strength of the ACK frame. The value of is "-1" in case
is equal to "NOACK". • is the transmission speed. In this case, "540" rapresented the fixed transmission speed equal to 54 Mbps. The files list_train_ch1.txt, list_train_ch5.txt, list_train_ch9.txt, list_train_ch13.txt contain the list of dataset segments used to perform the train for channels 1, 5, 9, 13, respectively. Instead, files list_test_ch1.txt, list_test_ch5.txt, list_test_ch9.txt, list_test_ch13.txt have the same meaning, but for the test. In total, the dataset consists of slightly more than 1142 hours for the train (corresponding to about 48 days) and slightly more than 642 hours for the test (corresponding to about 27 days). [1] G. Cena, S. Scanzio and A. Valenzano, “SDMAC: A Software-Defined MAC for Wi-Fi to Ease Implementation of Soft Real-time Applications,” in IEEE Transactions on Industrial Informatics, vol. 15, no. 6, pp. 3143-3154, June 2019. doi: 10.1109/TII.2018.2873205