Welcome! Systems and Automation

Overview defining automation and reviewing design considerations for automated systems.

Google Self-Driving Car, 2016

Hello, and welcome to my blog! Over the next ten weeks, I will be posting about a range of topics centered around the theme of human-machine system design. More specifically,

I will be discussing design challenges that arise from the complex dynamics of human experiences with systems and machines.

This draws in on human-centered design and how we consider the role of the user as we develop systems of the future. Enjoy!

What is Automation?

When you hear the words 'automation' and 'autonomy,' what pops into your mind, and how do you picture yourself fitting into the situation? Do you envision yourself sitting in retirement as a robot completes your daily chores? Or do you see yourself asking Siri to set a reminder for you to pick up cereal when you are at the grocery store? Maybe you imagine both, or maybe you imagine neither... Either way, you should note that there is a distinction between the terms 'automation' and 'autonomy.'

In simplified definitions, automated systems can only function within limited conditions and requires human interaction and supervision to achieve the desired task. On the other hand, autonomous systems are self-sufficient in their environment and are capable of adapting to human feedback to achieve a common goal with the user (Read Kaber, 2017, for a more detailed explanation).

Note that automated systems exist upon a continuum, ranging from minimal-to-no automation to a fully autonomous system. Take cars as an example. According to SAE International (Society of Automotive Engineers), there are six levels of vehicle automation ranging from 0–5:

• Level 0: no automation

• Level 1: automated driver assistance systems (e.g. cruise control)

• Level 2: partial vehicle automation (e.g. automated lane-keeping systems)

• Level 3: conditional automation (driver is not in primary control but is still responsible for intervening if needed)

• Level 4: high automation (all driving systems are automated, but a human can intervene and take control as desired)

• Level 5: full automation (all driving tasks are automated, no human intervention)

Due to these varying degrees of automation, the role of the user must also vary when considering the design of automated and autonomous systems.

Design Considerations for Automated Systems

The spectrum of automation leads to interesting challenges when designing automated systems. Some challenges to consider:

• What are you automating, and why? To increase the limits of tasks to which humans can accomplish? Because it’s cheaper and more productive than humans? In essence, what is your purpose?

• What is your desired span of automation? Think: the extent of system control your automation will cover and the time period your automation acts over.

• What are the potential consequences of a failed automation system, and how can you avoid failure? What checks need to be in place?

• How reliable is your system, and how much trust do your users place in the product?

While all of the above design questions are incredibly important, I believe the perceived reliability and correct levels of trust in your product by users is the most important from an interaction standpoint on the success of your product.

• Perceived reliability: this boils down to ensuring that the user understands the limits and scope of your automation—if misunderstood, the user may believe your product is unreliable due to it's inability to perform out-of-scope tasks.

• Correct levels of trust: the last thing you want is for users to place an over-abundance of trust in your automation, only to become inattentive and unable to intervene as required. Conversely, you don't want the product to have so many alerts that the user is desensitized to all notifications due to alert fatigue.