Robots in a dairy farm

I saw this video from Bloomberg about the use of robots in dairy farms. It's always a challenge to work in an environment that constantly changes such as a farm. However, that didn't stop the company in the video to create 3 robots for tasks that used to be the domain of human workers:

• Feed moving
• Dung cleaning
• Milking

Of the three tasks above, I consider milking the most challenging as it involves a moving creature with non-rigid body. Issues such as force control (to prevent cow discomfort or injuries) and dynamic visual servoing (to catch the moving udders) are some of the control issues that come to mind. Other engineering problems would be the robustness of the robots, battery life or power (for the mobile robots), and cost.

The robotic solutions doesn't come cheap (USD 200K per robot).

Maybe full automation is not really necessary? If the goal is to increase productivity, would providing farm workers with tools that could assist in executing his job at high speed be more cost-effective?

The irony of social media -- as pondered by Simon & Garfunkel 50 years earlier

A video is going around social media titled, "Look Up". The video essay speaks of how social media (mobile phone?) actually makes people unsocial. With eyes glued to mobile phones, opportunities to connect to one another just pass us by. It's a call to pry ourselves away from our phones, hence, the title.

Half a century ago in 1964 the duo Simon and Garfunkel, through their song The Sound of Silence also warned of technology that could cause isolation -- television.

I learned about the real meaning of this song in a poetry class in university,  32 years after the song was released. I couldn't relate to the song because I grew up in a generation when watching television is a norm (even necessary for some as it is one of the cheapest forms of entertainment). It felt like what the singing duo warned about didn't really take place or maybe it did, but a fish wouldn't really know what water is.

Humans in general would adapt to any given situation. Whether it's for the worse or the better is a matter left to future generations to find out. As for mobile phones, I think we just have to remember, "too much of something is bad enough".

Journal Impact Factor (or pissing contest for academics)

 

The comic strip above from PhD comics is a very humorous take on impact factor (IF) (although I believe most of the items included in the equation are TRUE).

Elsevier has recently released new metrics for journal impact factor. Three new computations are being introduced as better methods of assessing the importance of the journals and published articles.

The three rankings are:

• SNIP - Source Normalized Impact per Paper
• SJR - SCImago Journal Rank
• IPP - Impact per Publication

Universities and funding agencies reward researchers who get published on Journals with very high impact factors, such as Cell, Nature, Science, Molecular Cell, Nature Genetics, Lancet. Getting published in any of these journals signals that you have arrived as a researcher, which of course translates to opportunities (grants, seats on prestigious committees or boards, consultancies).

Exactly because of the importance of impact factor (IF) that a lot of researchers have also looked at the way IF was being computed. The three ranking methodologies above are a move towards a more realistic assessment of the actual value of the journals and published articles.

My only complaint in all of this as a researcher is the absence of engineering journals at the top of the list. Based on SJR's website, the science journal with the highest SJR is Annual Review of Immunology (SJR = 30,095), for mechanical engineering it's International Journal of Plasticity (SJR = 3,675), for electrical and electronics engineering it's IEEE Transactions on Pattern Analysis and Machine Intelligence (SJR = 8,094).

When I have the time in the future, I would take a close look at the various ranking computation and try to figure out why engineering journals rank very low on the impact factor ladder.