TI Training home
Wireless network challenges and solutions for a smarter grid IoT
System-level examples for wireless networks on smart grid IoT Sub-1GHz sensor to cloud industrial IoT gateway reference design system performance and demo
Wireless network challenges and solutions for a smarter grid IoT
2.2 Sub-1GHz sensor to cloud industrial IoT gateway reference design system performance and demo
Hello, everyone. This is the second training session of sub-1GHz Sensor to Cloud Industrial IoT Gateway Reference Design. In this training session, we will look into more details on the TI 15.4 MAC stacks, which is the underlying protocols for IoT gateway.
OK, this slide shows the details on how frequency hopping is operating with an example. The first row shows the frequency hopping pattern over time. In this example, we assume that the three channels are used for frequency hopping.
The second and the third row shows the packet transmissions over time from gateway and the sensor node respectively. The color of the box indicate which channel is used for the transmission.
The very first activity from the sensor node is network discovery. To discover the gateway or master node to be joined, the first action by the sensor node is send a PAS packets. The PAS stands for PAN Advertisement Solicit message. Since the sensor node don't know about which channel the gateway is listening and running.
And so in the example, it started with the channel number 16 and then wait for a certain amount of the time and then switch it to the next channel of the channel 8. But in this example, there are no example, because gateway is not running on those frequency.
And then finally gets response from gateway with a PAN advertisement, which is a PA message over channel 2. So with the PA, the sensor node knows the hopping pattern and the timing for unicast data.
And then the sensor node stays with the channel 2 to complete the remaining discovery process by sending out the PAN configuration solicit, which is a PCS. As a response, gateway send the PAN configuration back to the sensor node, which contains a broadcaster channel hopping pattern and a security key.
With these exchanges, the sensor node starts tracking the channel hopping pattern with gateway and ready to move to the next step of the joining process.
For joining process, it follows the 802.15.4 based association process. As a result of the joining process, the sensor will obtain the short address and PAN ID to be used as a unique identifier in the network.
Sensor node then start to data transmission based on the channel where the gateway is residing.
One thing I would like to note is the indirect transmission feature here, which can save power consumption when direct and bi-direction communications are required. Especially it is matter of the data communication from gateway to the sensor node.
The basic idea is that instead of the RX on always to get the gateway data, the sensor node poll to the gateway to see if gateway has data to send to me or not. Other than this period of time, the sensor node can turn off RX.
And then the last two data transaction on the channel 2 shows these indirect transmissions. So the sensor node polls gateway and [INAUDIBLE] with indication of pending data. When the sensor node receives this error, continue to enable the RX until getting data of data from the gateway.
This slide and the next slide show performance analysis. Here we show the packet delivery ratio with frequency hopping. The test scenario is given here. Interference was injected on channel 8. And the frequency hopping operates on the three channels, channel 8, 16, and 24. And we assume that the interference is injected on the channel 8.
The table here shows the packet delivery ratio. This shows how robust frequency hopping is compared to the non-frequency hopping mode. When about 2% of duty cycle interference is injected to the channel 8, non-frequency hopping, of course, have a significant impact on performance by lowering to 22% of the delivery ratio.
On the other hand, since frequency hopping is not staying always on the interference channel, the frequency hopping is not affected by the interference. So it is still achieving the 100% delivery ratio.
A note here is that our TI 15.4 MAC stack is built with some intelligent algorithm to detect and avoid transmissions on interference in addition to the frequency hopping. So this is why in the even 2% of duty cycle interference is there, so the popping pattern visiting channel 8 anyway, but still having the 100% packet delivery ratio.
Then even with 100% duty cycle interference on a single channel and the two channel cases, non-frequency hopping does not working at all since it's 100% jamming. So is no way to communicating on that channel. While frequency hopping still performing great with only the minor proponents degradations.
This slide shows the total current consumption analysis with the indirect transmissions. In the previous slide, we've discussed the indirect transmission, which is polling to the gateway to extract data instead of always turning on to receive data. So this slide shows how much we can improve the power consumption performance with the indirect transmission mechanism.
For the analysis, we assume that application receive 60 byte data over 50 Kbps FSK channel in a minute. And then the RX and the TX current consumption is based on the data sheet for the CC1310.
The first scenario is turning on the RF always. And then the second scenario is indirect transmission, so that it only turning on the RX when polling it. So for the indirect transmission, we assume that the polling interval is 10 seconds. Based on the result in the table shown here, we observe that there is a huge gap in the total current consumption between two scenario.
For the first scenario, always turning on the RF, the total consumption for one minute is about 324 millicoulombs. But the indirect transmission mechanism case the current consumption is only 0.5 millicoulombs.
You might concern about the timely response with the indirect transmissions. But this can be configured with the polling interval easily if you know about the maximum delay for the response. Of course, the power saving efficiency will vary, vary with the polling interval and the application profile.
OK, now is demo time. This video shows a sensor to cloud demo.
Hi, everyone. This video shows the sensor to cloud demo. This demo runs a two-node setup. One is a IoT gateway. Another one is a sensor node.
Let's first take a look at the IoT gateway. IoT gateway has two EVM here. One is a [INAUDIBLE] running the IoT agent on top of the Linux kernel. And then CC1310 LaunchPad running the sub-1GHz RF and the IEEE 802.15.4 [INAUDIBLE] TIMAC. And the UART interface is-- UART driver is running to connect it to the [INAUDIBLE].
And then [INAUDIBLE] has in addition to the RF connectivity, it also have a internet connectivity to access to the internet.
And then let's move on to sensor node here. Sensor node is running on the CC1310 standalone node. These CC1310 running the same protocols, the sub-1GHz RF and the TIMAC. And also to run the demo, it runs a sensor predications as well.
So now let's move on the monitor from the PC that can access to the internet. So from the PC, I type the web address to access to the cloud. It shows the IoT dashboard. And then it has the network information and also the network connectivity topology graph. And on the bottom side, that you can see the sensor node information here.
For this specific demo, it is collecting the temperature information every 10 seconds, as shown here. And in addition to the monitoring, this one also can do some simple-- also showing some simple examples of the controlling. There is a [INAUDIBLE] button here. This is controlling to the sensor node LEDs.
So let's take a look at the sensor node here. So by just clicking the button, and then what happening is that these LED is changing like this.
And as you know, this is also the cloud demo, which means that any device that can access to the internet can just taking the data, collecting data or control data from the sensor node.
So here is an example. So I have a cellular here. And then I just type the same web address from the cell phone. And then we can get exactly the same information, the same monitoring and controlling from the cellular phone.
Thank you for watching the video.
OK, this conclude the sub-1GHz Sensor to Cloud Industrial IoT Gateway Reference Design part two training session. Thank you for watching.
2017年 4月 10日
This training will cover system- and software-level deep-dives on key communication protocols which comprise a 6LoWPAN-based RF network. The training will review new TI Designs/products for an RF network implementation on the CC1310 wireless MCU and Beaglebone Black (BBB)– including both the end node and edge router (or data concentrator) solutions.