Pestology Blog

Feedback Control of Automatic Navigation for Cyborg Cockroach Without External Motion Capture System

We mostly think of cockroaches as a nuisance- they’re dirty, hang out in gross things, eat our foods, and have the potential to spread diseases. A recent meta-analysis of the contaminants that cockroaches carry measured in at least 78 different species of bacteria.* But, in perhaps the ultimate villain to hero storyline, what if we could use these ordinary pests as something extraordinary in times of crisis?  

Enter, the cyborg cockroach. Half computer, half roach, this combination of nature and machine has the potential to be a game changer for search and rescue missions in urban areas. Cockroaches are well-suited to search and rescue missions, as they are small enough to crawl through small openings in rubble, can self-right themselves if they get overturned, are adept at moving in dark and sheltered places, and are great climbers.  

Previous studies concerning cockroaches in search and rescue missions mainly focused on using cockroaches as tiny videographers, where cameras were attached to battery-powered “backpacks” that the cockroaches would wear. However, this method left a lot to be desired, as these systems take up a lot of battery power, require a lot of human input, and don’t always provide the best images in darkened areas. In addition, cockroaches tend to stick in one place once they reach an area that they deem appropriate shelter, like a dark corner. Therefore, the purpose of this study was to examine the potential of cyborg cockroaches operating and navigating complex environments “on their own” thanks to machine learning.  

To prevent cockroaches from slacking on their important search and rescue mission jobs, they were each fitted with an individual wireless stimulator backpack. The backpack features multiple tools and sensors for the job. For one, it acts as the stimulant for the cockroach to keep exploring disaster zones by stimulating the antennae and the cerci. Both are important sensory organs for cockroaches that help them navigate their worlds. Second, it contains a transmitter that wirelessly sends data about the location and movement of the cockroach back to the computer. This allows the computer to determine where the cockroach should go next. Third, it is fitted with a tiny laser sensor that can help determine the distance of the cockroach to a solid object. Lastly, it’s also fitted with a small thermal infrared sensor to detect human body heat. All of this is packed in a backpack that is 0.8 inches long on top of adult Madagascar Hissing Cockroaches.

The researchers conducted experiments where the cyborg cockroaches were asked to navigate difficult areas, such as sharp corners, and whether they were able to find a human in a complex environment. The cyborg cockroaches had a success rate of 82.3% navigating sharp corners in a complex environment, which is a huge win for a species that generally likes to stay hunkered down in those areas. The researchers also were able to create a user interface that allowed humans to “take control” of the cyborg cockroach and manually operate the cockroach within a complex environment with the laser and infrared sensors as guides, rather than cameras. However, most importantly, they were able to successfully conduct a trial where a human was detected in a complex environment by a cyborg cockroach that was completely run on machine learning without any sort of human intervention. While this human detection occurred at a short distance of 110 cm, the experiment still emphasizes how relevant and important these cyborg cockroaches could be. To view the cyborg cockroach in action, click here: https://drive.google.com/file/d/1plXFI0Isn6a8Bs3fHcDmFtb37AJ9WEcE/view ) 

This on-board obstacle avoidance packed into the cyborg cockroach backpacks has the potential to be a game-changer. While there’s a lot of testing that still needs to be done, the fact that a completely automated cockroach was able to successfully find a human in a testing scenario is a great first step in a tool that could potentially be used to save lives.  

In addition, this kind of technology has a lot of potential in the pest management sphere. Cockroaches often engage in aggregation behavior. Potentially, the use of different species of cyborg cockroaches could lead you directly to the source of the infestation. Cockroaches are much better at finding each other than we are, so the use of cyborg cockroaches to tell you exactly what areas you should be addressing with your pest management tools could not only save resources, but also time taken for inspection.  

With all the beneficial potential for this kind of technology, I, for one, accept our new cockroach robot overlords.  

Article by Laura Rosenwald, BCE

References

Ariyanto et al. 2024, Osaka University (Japan) and Diponegoro University (Indonesia), Published February 29 2024. Feedback control of automatic navigation for cyborg cockroach without external motion capture system 

*Nasirian, H. 2019. Contamination of Cockroaches (Insecta: Blattaria) by Medically Important Bacteriae: A Systematic Review and Meta-analysis. Journal of Medical Entomology, Volume 56, Issue 6. 1534-1554.  

Distinct Communities under the Snow Describing Characteristics of Subnivium Arthropod Communities

A common misconception is that arthropod life simply dies off in the colder months. In reality, we know that arthropod life survives in some capacity all year round thanks to a pretty cool list of ways that they can survive sub-freezing temperatures. Many organisms can produce antifreeze-like substances that prevent their cells from freezing, while others may use various methos of cold avoidance such as finding a warm place to hide to wait out the winter months. A lot of arthropods may use a combination of both strategies, antifreeze-like stuff to keep their bodies from freezing while seeking protection from harsh conditions under leaflitter or even under the snow itself.

Surprisingly, snow can act as an excellent insulator and provide a blanket of protection underneath. Specifically, when the snowpack reaches about 15-20cm, it can create a stable environment below that can remain warmer than ambient air temperatures exposed to freezing winds and other conditions. This protective region is known as the subnivium and can serve as a winter refuge for a long list of organisms.

While it’s known that life does survive in the subnivium, there are a lot of questions surrounding the diversity and abundance of arthropods that can be found in this insulated zone. To address this gap in knowledge, a group of researchers at the University of New Hampshire surveyed arthropod communities in a remote forested area using pitfall traps. They collected samples over two summers and one winter to compare what they found across seasons.

In total, over 20,000 arthropods were collected including various species of collembola, mites, spiders, beetles, millipedes, and centipedes. Collembola and mites were the first and second most trapped arthropods. Not surprisingly, samples collected during the summer months accounted for approximately 6-fold more arthropods than winter collections. The researchers also reported that insects accounted for roughly 18% of collection in winter and 25% in summer months. And, that about 95% of everything collected were adults.

As predicted, communities in winter months tended to be lower in most measures of abundance and diversity compared to the summer collections. Specifically, the subnivium community was much lower in overall abundance as well as biomass. In other words, the range of different species and the overall total number of bugs they caught in the winter months was lower than the summer months. What was really interesting was that some arthropods were dominant in the subnivium but rare or absent in the summer collections. About 37% of species were exclusively recovered in the winter, with rove beetles and spiders being the most prominent insects recovered in the winter samples (~26.6%).

This paper aligns with other past studies demonstrating that life simply doesn’t pause during the winter months. There are a wide range of diverse and distinct communities of arthropods that have evolved as specialists to not only survive but THRIVE in subfreezing temperatures under snowpack. This research also shows that, while the arthropods recovered in this study may not be high on the structural pest control’s most wanted list, heavy snowfall doesn’t equate to a decline in arthropod populations. And one could argue that heavy snowfall would even help to protect some arthropod communities (like ticks) by providing a thicker layer of insulation that shelters them from subfreezing air temperatures.  

It is important to point out that the way in which samples were collected does inherently have some level of trap-bias that affects what would be recovered. By that I mean a pitfall trap may not be the best way to survey for ticks or other organisms that we know to be active and capable of surviving the winter months. Having said that, this was a really cool study that added to our knowledge and understanding of what these winter communities may look like across seasons.

Article by Mike Bentley, PhD, BCE

References

Christopher P Ziadeh, Shayleigh B Ziadeh, Breanne H Aflague, Mark A Townley, Matthew P Ayres, Alexandra R Contosta, Jeff R Garnas, Distinct communities under the snow: describing characteristics of subnivium arthropod communities, Environmental Entomology, 2024;, nvae017, https://doi.org/10.1093/ee/nvae017

Ants Evade Harmful Food by Active Abandonment

This article looked at the Argentine ant which is invasive in the US but native to Argentina where this study was conducted. The idea here was to determine if ants abandon toxic bait, a commonly used product in ant control.

The researchers created a sugar solution that has been known to be well accepted by these ants. The toxic solution was made using boric acid as the active ingredient added to that same sugar solution. An important factor in this strategy is that mortality is delayed as is critical for ant baiting to return to the colony. They tested this out in some pre trials, and the ants lived for at least 6 hours after consumption. They also confirmed that the bait being used was palatable to avoid the possibility of distaste rather than avoidance behaviorally.  

The researchers ran two main experiments in this study. The first was day-scale dynamics and spatial extent of trail abandonment and the second was hour scale dynamics of trail abandonment so that's a fancy way of saying they looked at smaller and larger time scales.

To begin each experiment, they started by luring the ants into a false sense of security by providing a sugar only solution that was not toxic over a bridge to a set of feeders as seen in the diagram above. They placed this structure right by a known trail of ants that were heading to the nest which is called the trunk trail.  

Data collection consisted of measuring ant activity which was the average number of ants crossing the bridge per minute averaged from 3 minutes.  They also measured ant activity on the trunk trail to make sure reduced activity wasn’t a result of the colony dying off.  

For the first experiment, they had two bridge setups 7 meters apart from each other along the trunk trail. They took activity measurements at one hour intervals three times that afternoon. Over the next two days, they took measurements during the mornings and afternoons getting average activity levels for each day’s more active hours. They did this whole thing 5 times on different trunk trails as repetitions.  

This showed that trail abandonment occurred rapidly within about 18 hours of the toxic bait being provided. Fewer ants were found going to the toxic bait at each successive time interval while numbers at the sugar only station remained similar. This means that despite the bait still being palatable, fewer ants went after it as time went on and it was not because they colony had been killed off. 

Based on this result, they continued on to look at a closer temporal scale. To refine the experiment, they had a similar setup with just a few differences. Once the toxic bait was swapped in, the activity measurements were taken every hour from 10am to 5pm. They did all this 6 times on different trunk trails.  

What they found with this closer look was the ant foraging at the toxic bait decreased drastically over just three hours, with a 43% reduction in ants per minute. This increased to 79% reduction by 6 hours. I want to point out that this is all while the regular sucrose bridge did not see a decrease in numbers which indicates that “the ants were not satiated, not killed, and continued actively foraging throughout the experiment.”

With all that said and established, how does this impact using baits to control ants? The 3 hour time frame before foraging was reduced becomes a critical time frame for initial bait placement. Ideal bait placement is important so that as many ants as possible consume the bait before they begin to abandon it. Placing many baits can increase likelihood of more ants running into it.  

The researchers suggest that palatability being separate from abandonment could mean that other ant baits previously thought to be unpalatable were actually just being abandoned and ultimately, adjustments to such baits could be made, making more available products to use.  These experiments were also only tested on one active ingredient so it will be important for such work to be replicated with other common active ingredients.

Article by Ellie Lane

References

Zanola, D., Czaczkes, T.J. & Josens, R. Ants evade harmful food by active abandonment. Commun Biol 7, 84 (2024). https://doi.org/10.1038/s42003-023-05729-7