Autonomous robotics for sustainable agriculture. Based in the south of the UK. Prototypes of an autonomous mechanical farm-scale weeding robot currently beginning real-world testing. Still a huge amount of work to do though.
Hardware and software developed fairly much from scratch, not using ROS (for not entirely crazy reasons...); everything written in Rust which I find well suited to this application area.
The robot is built using off-the-shelf components and 3d-printed custom parts, so build cost is surprisingly low, and iterations are fast (well, for hardware dev).
On robot compute is a couple of Raspberry Pi 5s.
Currently using the RPi AI Kit for image recognition, ie Hailo 8[L] accelerators.
Not currently using any advanced robotics VLA-type AI models, but soon looking to experiment with some of it, initially in simulation.
Feel free to get in touch if you'd like to talk :) Contact details in my HN profile, and on our website.
I invested in Small Robots Co (https://smallrobotco.com/) who were doing something similar. They had a good product, good brand, raised some funding, and were starting to get traction, with robots in trials on a number of farms, but at the end of the day they ran out of funding before they reached default-alive.
It's a tough space - convincing farmers to give it a go and running trials takes time and the UK isn't a very startup-friendly environment - investors are too often looking for a quick return.
This is such an important area - it's only going to become more critical to be able to grow more food whilst using less fertiliser and weed killer - so I wish you the very best of luck!
Coincidentally, we're based fairly close to where they operated. We are in touch with some of the people that used to be involved with SRC, and have been able to learn from some of their experiences. There is agreement that the UK can be difficult for this kind of startup, but also about the importance of the product area.
Before we started building we considered many different designs, including legs. However, it introduces significant extra mechanical and control complexity, with more complex failure modes - e.g. one leg gets stuck in the mud; it also would be more expensive to build.
So we decided to stick with wheels, at least for this product iteration!
My knowledge of ROS is a couple of years out of date, but primarily that reproducible testing and simulation, with precise time/clock management, which is essential for a reliable product, was very difficult in ROS.
I also felt the ROS build system more convoluted than necessary; and seemed rather brittle - it was too easy to break it with OS or other updates.
We found that many off-the-shelf ROS nodes didn't do quite what we wanted, and ended up spending much more time than expected rewriting code that we expected we wouldn't need to. It is quicker, and we end up with less & more maintainable code, by writing it ourselves.
I expect this could have been resolved, but when testing ROS we also ended up using more compute resources on-robot than we expected.
Using our own system allows us to build exactly what we require, which has become more important as our system gets larger and more complex; and means integrations into other systems (including testing) are easier.
You are already the second case in two weeks that have abandoned ROS for industrial purpose(and not university), preferring to build something of their own.
I agree that the build system is more complicated than it should be,but I was unaware of the problems related to the resources used by the nodes.
Your comment gave me a lot to think about, thank you.
These are apparently not large EV-type rechargeable lithium batteries, as I immediately assumed.
From the article: "Aricell makes lithium primary batteries for sensors and radio communication devices". A "primary battery" is non-rechargeable; and given the use-cases mentioned I expect each individual battery is fairly small.
Of course, when there are 10s of 1000s of them together that's still a lot of energy to burn.
Primary, non-rechargeable lithium batteries typically contain metal lithium [0], and are actually more likely to catch fire when mishandled, compared to Li-ion stuff.
Specifically, most common Li-ion fires start when overcharged (especially with high current and in cold), and from short-circuits (e.g. when pierced). But only have a very small chance of spontaneously igniting from just disassembly alone [1]. Still non-zero chance, don't open them!
Primary/metal lithium batteries, on the other hand, are much more likely to burst into flames when opened. Notably, lithium-iron disulfide (AA/AAA "alkaline replacement") cells are notorious to do that just from air exposure, even if one is very careful to not short/pierce anything.
I'm using latest FF, it did work, I said I was an idiot because despite seeing it work, trying it out in a non-Safari browser, I came here and wondered aloud if it would work other than in Safari!
Agreed — I was worried the prequel would be a bit of a money grab follow-up on the first one's popularity, but (1) it's a standalone story which is great in its own right, and (2) in the few places where it does have some narrative connections with the original, it is done very well, and tragic/bittersweet.
Yes, I am always worried about prequels for that reason, but I think I enjoyed this one more than the original. They are really very different in scope and story as you said, and the bittersweet quality of the connection between the two books enhances the ending of the prequel in my view.
1. This article is specific to the UK, not the US.
2. The article is published by an Electrical company, obviously as PR piece to make it look like they have things under controll. This is not an idependaty study.
3. One the "solutions" to the problem the UK came up with is regulations which state:
>The regulations ensure charge points have smart functionality, allowing the charging of an electric vehicle when there is less demand on the grid, or when more renewable electricity is available. The regulations also ensure that charge points meet certain device-level requirements, enabling a minimum level of access, security and information for consumers.
In other words, even the grid the UK cannot handle an influx of EV's to combat this they are making the charge points "smart" (i.e. you won't be able to charge your vehicle when you wish only when the "smart" point says demand is low enough). This is hardly a solution. It attempts to "spread" out the charging with these "smart points", the problem is most people work during the day (and are parked somewhere where they cannot charge) and will want to charge at night. This would leave many SOL when they wake up to find their vehicle hasn't charged at all.
Hard, fucking, pass. Additionally, EV's are complete non-starter until they can go from 0% charge to full charge in 2 min or less IMO. If am on a road trip and trying to make good time I am not waiting around an hour+ for my vehicle to charge. Until charging is as easy and quick as pumping gas I have no interest in an EV.
EV’s draw less power on average than space heaters, and rapid charging is bad for the battery. These two facts address most concerns about grid overload.
On this road trip, you only stop for 2 minutes every 300-600 miles. Are you peeing in a bottle or something? Current cars take 20-40 minutes for 20-80%. That’s more than fast enough if you stop for meals.
Also, charging is more convenient than pumping gas, since 99% of the time, you do it at your destination (home, work, restaurant, grocery store, etc.)
when it said "our electric grid is nowhere near where it needs to be to remotely support a large shift to EVs"
without specifying further detail (like "we need more chargers connected to that grid"), the reasonable interpretation is that the grid itself, meaning the means of delivering power to power users (including any installed chargers), has deficiencies preventing it from doing so
I own a Peugot iOn (rebadged Mitsubishi i-MiEV) manufactured in 2012, so it's 9 years old. I bought it second hand, and the service record shows it's needed zero maintenance (aside from tires). I have not needed to perform any maintenance in the ~2 years I've owned it.
I'm not in any way recommending this specific car (although it suits me perfectly); but it shows that even a fairly early electric car like this is 10-year capable. The only downside has been the expected gradual decrease in range.
I've also owned a Tesla Model 3 for 16 months, and it's been maintenance and trouble free. I'm not denying that some people have problems with them, however that's not been my experience, nor the experience of the two other people I know who own one.
I probably plan to keep the Tesla for many years, so hope my trouble-free motoring continues :)
The energy density really is abysmal, which is why the "lifting heavy things up" techinique of energy storage is essentially never discussed outside of pumped-storage.
The problem is the linear relationship between the mass, the height, and the stored energy: energy = mass * height * gravitational-acceleration.
gravitational-acceleration is fixed at the earth's surface to ~10m/s2
So taking an example of 1,000,000 tons lifted up 100 meters:
This looks like a lot, but really isn't. It's equal to ~278 MWh (megawatt hours), which means it can supply 278 MWs for one hour. 278 MWs is equivalent to one small power station.
Note that the largest pumped-storage power station in the UK, which is of course constrained by exactly the same E = mgh formula, Dinorwig (https://en.wikipedia.org/wiki/Dinorwig_Power_Station) stores ~9,000 MWh.
Another way to consider this is to calculate how much mass needs lifting 100m to supply the whole of a country for a day.
As a very crude estimate the UK requires an average of about 30,000MW of electrical energy. Over a day this equals 30,000,000,000 * 24 3,600,000 Joules = 2.510^18 Joules per day.
The mass required to be lifted up 100m to store this is 2.510^18 / (100 10) = 2.510^15 Kg = 2.510^12 tons = 2,500,000,000,000 tons.
I live in an extremely hilly area, made of soft rock (limestone), but very little water. Carving out a hill to 100m down, with a 100x100m cross section is totally doable—we have larger projects just to make flat land to build houses on. If the weight was only 10m high, it’d be ~200k tons. By your calculation that’s ~54MWh/day, or enough energy to power ~1500 homes. There’s a hill the right size next to my village of ~1200 homes. We have excess wind & solar power.
But what kind of machine do you need to raise a hill up and down by a few meters ? Even if breaking the load in small bits, the cost and maintenance of equipment would likely be a few magnitude higher that the energy stored or saved.
An electric motor moves it up. An electric regenerating break (an electric generator) keeps it from going down too quickly and/or not at the right time. The particulars of maintenance and long term robustness are pretty straightforward engineering efforts. The largest steam hammers are 125 short tons, and they can operate for decades.
Why would the stresses be any greater than burying a turbine electric generator at the bottom of a hydroelectric dam? The forces would be similar, right? That's the whole point: it's just a crap load of "pressure" due to a bunch of stuff piled up on top.
I presume another application could be higher density electronics owing to reduced heat generation owing to reduced dissipation. However, if I recall comments here correctly we are getting to the point where electromagnetic shielding may be required as we scale down further which provides a second limiting factor to increased densities.
PS. Last year ETH Zurich also looked at weaving nanothreads in a kagome pattern, https://www.ethz.ch/en/news-and-events/eth-news/news/2017/08... ... note also that this style of weaving is widespread across all of continental Southeast Asia, at least southwest China, Vietnam, Philippines, Laos, Thailand, Malaysia, Myanmar and probably further west in South Asia. It is real a shame that a lot of terminology falls in to the "some random American saw a Japanese name for something so it is termed Japanese in English" category (eg. various food ingredients, philosophical concepts, art history, etc.). People could learn a lot more if they had broader regional comprehension of Asia and its history.
It isn't like those sorts of inaccuracies are unique to Asia. "Danish" pastries are called "Vienna bread" in Denmark because they weren't introduced to Denmark until an influx of foreign bakers caused by the Danish baker's union going on strike.
Autonomous robotics for sustainable agriculture. Based in the south of the UK. Prototypes of an autonomous mechanical farm-scale weeding robot currently beginning real-world testing. Still a huge amount of work to do though.
Hardware and software developed fairly much from scratch, not using ROS (for not entirely crazy reasons...); everything written in Rust which I find well suited to this application area.
The robot is built using off-the-shelf components and 3d-printed custom parts, so build cost is surprisingly low, and iterations are fast (well, for hardware dev).
On robot compute is a couple of Raspberry Pi 5s.
Currently using the RPi AI Kit for image recognition, ie Hailo 8[L] accelerators.
Not currently using any advanced robotics VLA-type AI models, but soon looking to experiment with some of it, initially in simulation.
Feel free to get in touch if you'd like to talk :) Contact details in my HN profile, and on our website.