Prof. James Gross
As mentioned in the introduction, we will begin this series by covering the random propagation effects that occur within a wireless environment. These effects can greatly impact signal reliability and are a significant hurdle for industrial wireless networks to overcome.
This entry provides an overview of the issue, explains why current wireless technologies have not yet implemented an effective counter, then introduces a groundbreaking propagation-countering technique developed over this past decade. We finish by introducing EchoRing and its potential as a commercial industrial grade wireless solution.
The basic process of wireless communications begins when a transmitter encodes an information packet (a collection of bits) into an electromagnetic wave. This process is known as modulation. Propagation effects begin as the wave is sent over the air, which impacts the signal’s reception. The two most important propagation effects are:
This effect occurs when one or more “copies” of the original transmission reaches the receiver. This is because, in most cases, the exact location of the receiver is unknown to the transmitter, therefore the transmission wave is emitted in a wide arc in all directions. This wave then inevitably encounters and bounces off various physical objects within the environment (machines, walls, human workers, etc.), thus creating multiple signal copies that reach the receiver at different times.
An operator (sender) transmits to a robot (receiver), which receives
multiple signal copies due to the signal bouncing around the environment
This other main propagation effect is more straightforward; electromagnetic waves continuously lose energy and become weaker the longer they travel from transmitter to receiver.
A typical path-loss of an electromagnetic wave
Multi-path propagation and attenuation are not just limited to wireless communications, either. For instance, when a sound (i.e. a song played on a Bluetooth speaker) is reflected off several hard surfaces before it reaches a destination (a listener), an echo is heard. Likewise, it is obvious that the further away a sound is from a listener – especially when objects are in the way – the quieter the sound will be perceived. The sound analogy only holds up to a certain point, however. Returning to the context of electromagnetic waves, when multi-path propagation and attenuation occur simultaneously (such as in busy factory floors) they create two further random effects that can significantly weaken wireless transmissions:
Given that a standard industrial shop floor spans a considerable distance (i.e. several tens of meters), mobile assets such as forklifts, AGVs, untethered production robots, human workers, etc. can easily get in the way of a wireless transmission path while going about their duties. Whenever this happens, a traditional point-to-point signal experiences a large drop in strength until the path becomes clear again. Given that the speed, timing and direction of these mobile assets are near-impossible to predict, it is also near-impossible to predict when and for how long signal drops will occur. This random drop effect is known as shadowing.
The second, but more subtle, random wireless effect is known as fading. The best way to understand fading is to imagine an electromagnetic wave as an oscillation with a specific shape. Whenever multiple waves arrive at a receiver simultaneously due to the above-mentioned multi-path propagation effect, the signal the receiver actually processes is a superposition (overlap) of all received waves laid on top of each other. This superposition can either boost or weaken the signal strength, depending on how closely the received frequency waves align.
As mentioned above, a destructive, misaligned superposition is primarily caused by signal waves arriving at different times due to multi-path propagation. Given that each signal typically has hundreds of different wave copies – each with its own unique travel distance – it is impossible to predict when a destructive superposition will occur.
Lastly, it is important to note that while fading occurs 1000 times faster than shadowing (within the range of about 10ms), it also degrades signal strength ten times more severely.
The Combined Effect of Shadowing and Fading
As explained under attenuation, only a tiny fraction of a transmission’s starting energy actually arrives at its receiver. Thankfully, information packets can still be delivered even if the energy decreases by a factor of 1 million (or even 1 billion). A successful transmission hinges not on a signal’s total strength, but whether its strength is at least 100 times higher than the network’s noise - the ambient energy within the propagation environment. This means that a signal-to-noise ratio (SNR) can tolerate a strength drop by a factor of 10 without losing a connection. Unfortunately, fading and shadowing effects combined can easily compromise signal strength to this degree or or even higher. Noise-induced signal drops are therefore a key hurdle to overcome before wireless networks can be implemented in time-critical industrial applications.
Current Wireless Reliability Solutions, and Why they are Incompatible with Industrial Setups
A common wireless reliability solution is to simply re-transmit a corrupted data packet as soon as it fails to deliver, or after a set timespan. This is known as Automatic Repeat Request (ARQ). ARQ functions very well in non-critical wireless communications and is a standard component in mobile networks. Industrial applications however require constant real-time latency, while in most cases ARQ recovery results in noticeable connection lag.
Another common reliability solution is to increase a network’s error correction, allowing information packets to be retrieved even when noise outpaces signal strength. The issue with employing this method for industrial purposes is that it requires larger information packets to be sent, resulting in longer transmission times and defeating the purpose of real-time latency in the first place.
The third traditional solution is to simply increase transmitter power and receiver sensitivity. This method is not viable either, since it can be very expensive in overhead costs and is still insufficient to cover the most severe signal drops.
As it stands, we must look to novel solutions in order to meet performance requirements for industrial wireless networks.
Cooperative Communications as a Solution for URLLC Wireless Networks
Over the past decade an innovative wireless transmission method known as cooperative communication has emerged, providing a breakthrough solution for ultra-reliable low latency wireless communications (URLLC).
In its most basic form, a cooperative communication setup consists of a transmitter, a receiver and at least one backup relay somewhere in range of both. The transmitter broadcasts copies of a data packet to both the receiver and the relay, whereupon the relay also transmits the packet to the receiver for redundancy. This process is simultaneous and instantaneous, so no real-time latency is sacrificed.
It is through this signal redundancy that shadowing effects are overcome; whenever a direct link is compromised, at least one alternate connection has already been established. Even a single relay installed in a commanding location within a propagation environment (such as a ceiling) dramatically increases a network's real-time reliability. This reliability is further increased with each relay node installed.
Cooperative communication takes things even further when countering fading effects. To reiterate, fading can either boost or degrade signal quality depending on the alignment of the incoming signal wavelengths. With several cooperative network nodes placed in range of both the receiver and transmitter, chances are high that some transmitter-relay and relay-receiver wave paths will be constructive (as illustrated in the diagram above). Combined with real-time channel data gathering and analysis, a network can be optimized to automatically select the strongest signal path to the receiver. This technique is known as multi-user diversity - the key factor to improve wireless reliability to the point where the strict time/performance requirements of industrial applications can be met.
To summarize, cooperative communication is an extremely effective countermeasure to shadowing and fading effects, and has proven to be the most promising network setup for industrial wireless applications. It is also very cost-effective when compared to other systems that feature multi-antennas or complex infrastructure. That said, the process requires both 1) effective network planning to ensure that enough relay nodes are in range of each transmitter and receiver, and 2) real-time channel status information gathering. This is where the EchoRing solution comes in.
EchoRing-Enabled Cooperative Communications
First, a brief introduction: EchoRing is a wireless communications technology built around a token-passing protocol. Token-passing (located within a Medium Access Control protocol) refers to a unique control packet or “token” being transmitted between network nodes. A station in an EchoRing network can only transmit payload data while holding the token itself, which ensures both a set network latency and deterministic control over channel access. These are topics we cover in detail in the next entry to this series, "Avoiding Collisions: Adding Determinism to My Network."
Returning to our main topic, EchoRing’s token-passing protocol is an excellent solution for cooperative communications. This is because each EchoRing token contains up-to-date network status information, including the strength of each node connection. The token-holder (sender) is therefore able to instantly determine the node with the best secondary connection to the receiver, assigning it as the relay.
EchoRing’s cooperative communications method unfolds in two steps:
1. The transmitter sends copies of the information packet to both the receiver and the relay (or “echo” station). As mentioned above, the packet is transmitted in all directions in a wide wave, therefore only one is required to reach both destinations simultaneously.
An operator (transmitter), an assembly robot (receiver) and an autonomous forklift (echo station)
2. If for whatever reason the packet fails reach the receiver, the echo station is instantly notified and relays the packet instead.
Transmission via the echo station, after a direct transmission has failed
It is through this cooperative system that EchoRing ensures a cable-like reliability of at most 1 packet loss per 1 million (10-6) and a recovery rate of at least 99.99999%, depending on the application and its propagation environment. Even in the least ideal conditions EchoRing still fulfills the reliability requirements of most industrial applications.
We hope this breakdown was informative! Please review our next post for a detailed breakdown on how EchoRing avoids collisions and latency issues through network determinism.