In previous materials, we reviewed codecs and protocols Bluetooth, which allows ordinary users to use headsets, fitness trackers, and even functions in the car. Today, we’re going to take a look at the improvements and new features in the new generation of Bluetooth 6.
Content
With each transition to a new generation, Bluetooth technical characteristics improve. This time, the overall speed and latency of data transmission have increased. This is achieved by introducing new methods and optimizing existing ones. Of course, they are all invisible to the average user.
| Bluetooth 4.2 | Bluetooth 5.4 | Bluetooth 6.0 | |
| Maximum speed (Mbps) | 1 | 2 | ∼3 |
| Bandwidth (Mbps) | 1 | 2 | ∼3 |
| Delays (ms) | 150-250 | 25-40 | ∼20 |
| Electricity consumption | High | Low | Lowest |
| Maximum distance (m) | 60 | 240 | 300 |
Here is a list of improvements in the new generation of Bluetooth 6.0 and we will try to analyze them in the following sections:
Many applications need to calculate the distance between devices. Previous versions of Bluetooth used the Path Lost Calculation method. It requires strength measurement (RSSI) and the exact value of the received signal level drop at the control distance (1 meter). Applying math and knowing physics, we can see the unfortunate graph below. This method is especially vulnerable to interference and environmental influences. Therefore, the use of this method is not very safe and reliable. For example, for keyless entry systems for cars. Therefore, it is necessary to create a new more sophisticated method.
To solve this problem, the Bluetooth SIG standardization group is introducing a new technology called Bluetooth Channel Sounding in the new version of Bluetooth 6.0. It includes two new methods for calculating distance: Phase Detection Ranging (Phase-Based Ranging, PBR) and Determination of Data Exchange Time (Round-Trip Timing, RTT). The new system must ensure safety and have an ultra-high accuracy of up to several centimeters in determining the distance between the two devices.
The PBR method uses the fundamental features of a radio signal: phase, frequency, and wavelength. Here is an example from the image below. We have two devices that are in contact with each other. The first one is the Initiator, and the second one is the Reflector.
The initiator transmits a signal at a given frequency f1 to the Reflector, which «reflects the wave» back. The initiator receives the first data and calculates the phase Pf1 of the received signal. After that, the procedure is repeated, but with a different frequency f2 we obtain new data on the phase Pf2. Now we can determine the distance between the devices in the formula using the difference between the frequencies (f1 – f2), phases (Pf1 – Pf2) and the speed of light. In practice, for better measurement accuracy, the calculation is performed more than twice.
There is a certain problem with the PBR method — Distance Ambiguity. According to the image above, the phase values start to repeat after a while at different distances between devices. This ambiguity depends on the difference between frequencies called Frequency Division. Bluetooth CS uses a value of 1 MHz, so the ambiguity does not occur up to about 150 meters. To overcome this disadvantage, the RTT method is used.
The Initiator sends a data packet to the Reflector and creates a variable that records the time the packet was transmitted, known as ToD (Time of Departure). After receiving the packet back from the Reflector, a new variable — ToA (Time of Arrival) is created. After that, we determine the final time T(A-D)then the measure of the distance traveled is calculated by multiplying T(A-D) by the speed of light and dividing by 2.
This method also has a certain problem — the data packet will arrive at the Reflector, but it still needs to receive, calculate, form a response, and transmit the new data back to the Initiator. And all this time can take more than a millisecond.
Bluetooth Channel Sounding guarantees accurate time values for the Reflector’s operation thanks to various ToD/ToA capture methods, and this data can be used for further accurate distance measurement.
The specifications state that both PBR and RTT methods do not necessarily work together because high accuracy requires more computation, which increases latency. Therefore, the standard allows the use of only one method at the developer’s choice.
Interestingly, an attack detection system has been added to Bluetooth Channel Sounding. It indicates the probability of an attack on devices at the program level. Of course, Bluetooth controllers themselves also have systems to counteract them.
Bluetooth communication uses a complex Link Layer (LL), which describes the states of each device: Standby, Initiating, Advertising, Connecting, Scanning, Isochronous Broadcast, Synchronization. These states are shown in the image below. Each of these layers has its own Filter Policies, which should filter the received data packets and send them to the layers above.
However, there are times when Device #1 (the Sender) is busy announcing packets to search for other devices, rather than working with the already connected Device #2 (the Receiver). This results in packet drops, delays, and slower connection speeds. That is why a new, more complex ADV_DECISION_IND declaration method has been developed to replace the previous options. This will allow the Sender to receive the necessary packets from the Receiver. Other devices with the correct Filtering Policy settings will not interfere with already connected devices. This should be especially helpful in an environment with a large number of Bluetooth devices.
One of the main features of Bluetooth LE (Low Energy) is a new way of thinking about device discovery. By transmitting small data packets at regular intervals, devices inform each other that they are in range or ready to connect. Detection is achieved by scanning the environment with a radio receiver on the main Bluetooth LE channels. It is quite simple and convenient.
Of course, scanning and visibility depends on many factors: distance, signal strength, external environment with obstacles, etc. Don’t forget that these packages generate the same reports within systems, so you need to filter duplicate messages.
Receiving notifications before and after filtering
However, we need to know exactly when the «devices disappear from the» field of view. If you use the standard method, you need to scan the environment again. The new Advertisers Monitoring method does not require all this ad filtering. It allows the host to receive only two ad packets: when the desired device has moved out of range of the transmitter and when it has moved back in. For this purpose, the signal strength parameter (RSSI) is used. This significantly increases computing optimization and device tracking efficiency. This is especially useful for audio devices (headsets or headphones), as it reduces power consumption and extends battery life. This method was invented specifically for use in Bluetooth LE Audio.
Bluetooth LE Audio uses what is known as isochronous communication to ensure that multiple streams of related data are transmitted and played back to the required receivers at the same time. For example, in a wireless headset, audio is first transmitted to the left channel and the corresponding earpiece, then to the right channel, but both are played simultaneously. In Bluetooth, the Isochronous Physical Channel and the Isochronous Physical Layer are responsible for this communication.
As you may recall in one of the of previous issuesaudio is converted into a digital format during sample creation. It is these samples that codecs compare in size. A set of samples in a codec is called a Frame. They have a fixed duration and have a certain number of samples, which is determined by the Sampling Rate. For example, one frame of 10 ms duration and 44.1 kHz frequency contains 441 samples.
Looking at the complex scheme of ISOAL (Isochronous Adaptation Layer) above, you immediately realize that such a process can not only lose frames during reception, processing, and transmission, but also increase delays. That’s why Unsegmented Frame Mode was introduced. Roughly speaking, packets between the data link and program layers are not divided (segmented) into smaller parts. This not only reduces latency, but also increases reliability by reducing the risk of losing transmitted data packets.
The functionality of the Bluetooth protocol is constantly expanding. That’s why it’s important to maintain a clear definition between devices. For example, Device #1 supports the A2DP profile, but does Device #2 support it? For this purpose, there is a variable and procedure called FeatureSet in the Link Layer. Device #1 sends a FeatureSet to Device #2 indicating a kind of table of its capabilities.
Conditional example of a FeatureSet with profiles
It soon became clear that 64 bits were not enough to cover the wide range of functions defined in the main Bluetooth 6 specification and for future standards. Therefore, it was necessary to significantly increase the size of the FeatureSet and divide it into addressable pages. Page 0 contains the first 64 bits according to the Bluetooth 6 specification. And then there are 10 more numbered pages, each of which has a size of 192 bits.
Channel layer with old and updated FeatureSet
The specification of the new standard clearly states that different values of transmission time intervals can be used. Typically, this value is 150 µs. Now it can be increased or decreased as needed. This improves the overall data throughput and makes the technology applicable:
Updated scheme of the Frame Space Update procedure
The new generation of Bluetooth has received many optimizations and improvements: a significant increase in distance measurement between devices, more stable music transmission, a foundation for supporting future generations, improved energy efficiency, and other useful features that are invisible to ordinary users. We hope that devices with the new standard will be available soon so that we can appreciate their capabilities.