Видео

That's great news, good luck with your project!

Hi Milton! That's a great course to be taking to get into robotics - absolutely. If you try some of the languages and maybe robot projects I talk about in the video, it would be a good start to that career.

@@mikelikesrobots Thanks, if I want to become a robotic engineer do I need lots of math I don't like math that much what should I do

You don't really need that much maths to be able to work in robotics. There are transform libraries and maths libraries already available for you to use. You need to understand what you're doing, but not necessarily the maths behind it, so you should be fine. If you're already doing science and technology, you're likely doing enough maths already!

@@mikelikesrobots okay thanks

You're welcome!

There's two parts. First, you need to add the target in your SSH config. If you're on Windows, this is probably C:/Users/<your username>/.ssh/config. Add an entry like this: Host rpi0 Hostname rpi0 User <your user> Then, make sure the VSCode SSH extension is installed, and the option should pop up. You can watch this video from the URL to see me do it for a jetbot - the process is the same. ruclips.net/video/pozA2Fc9kj4/видео.html

Glad to hear it helped!

Glad you found it useful! Do let me know if there's anything in particular you want to see.

​@@mikelikesrobots first of All , Happy New year . I would like to see the integration of AWS location services within a sensor Fleet coupled to Amplify Geo for User interface. Something else maybe you can depict how to reduce Costs through batch message using appropriate Aws services, I guess that kinesis can be useful in this scope. thanks .

Happy new year! Thank you for your suggestion - I'll add that to my list of video ideas :D

@@mikelikesrobots I am a curricula specialist , with Elec eng backg, and I am working a new AWS services based curricula dedicated for industry4.0. whole set of services are deployed for iOt monitoring and to anticipate measurements through Sagemaker and other ML predictive tools available on iot sitewise , diprecated iot analytics , managed grafana , twinmaker , grafana ..etc . I'm still discovering those tools though I'm AWS saa certified but services are versatile and it's not easy to harness what matters to a specific project. You are helping me a lot with your videos . Thanks 🙏

That's great! I hope you manage to learn all about robots. Yes, the advice does apply to drones, it's pretty much the same issues as the rest of robotics. I think it varies more for drones whether ROS is used or not, but learning those languages and practising using a real project are still great steps towards working on them full time.

I totally agree, even just providing precompiled binaries would make life a lot quicker. Hopefully this is something that AWS can provide in future.

Don't feel bad! Flood fills and recursion in general take a new way of thinking. For flood fills, try to think of how to test every possible point by calling the recursive function with the next point. From a point, check if it's the end, then get the neighbours to search from, and for each one call your function. Then think about how to join everything back up together if you do find a valid solution. Try writing a sudoku solver to practise the backtracking. The best thing you can do is try variations of flood-filling to practise. As far as correctly using the Rust tools, I honestly wouldn't worry too much. Get comfortable with the language first. Get used to iterators and functional style syntax, the borrow checker, and so on. Then you can look for ways to improve your code - ask people online to review it, read other code in open source. Choose one new thing at a time to incorporate into your code, and practise it. Advent of Code is GREAT for this - it's the primary reason I'm doing AoC! Finally, you can post your code to the rust subreddit and I'm sure people will happily give you pointers. If you can find a mentor to learn from, even better! Good luck!!

I'm glad you enjoyed it! That's exactly what I was hoping for with this series. Just a shame that in the later puzzles, it's too challenging to keep up the pace while using this style.

I would recommend robotics engineering. My personal opinion is that you can learn a lot of the programming from a computer science degree on the job. That doesn't mean it's not worth it - it's more a question of where is the best place to learn the concepts. I would learn programming at home and on the job, and the robotics at university. Good luck whichever you choose - they are both great options!

The most common in robotics are Python and C++. There may be others, depending on the course, but I would expect to see at least these two. Also, there's a lot of overlap between C++ and C, so you'd have a good head start on learning C as well!

You're welcome! Another commenter has pointed out the same thing, and it was such a "d'oh!" moment! It just shows how easy it is to assume something and then never question that assumption.

I'm not sure how to do that - are you saying during the dijkstra algorithm, you can track your current direction to apply the turn penalty? I couldn't figure out a way of encoding the current direction without that extra complexity or making a (Point, Direction) data type.

Glad you're enjoying the videos!!

@@santiagogamboa7387 This challenge was really difficult. Last video I got pretty frustrated trying to figure out an algorithm. I'm sorry to say that I don't have a lot more to tell you than the obvious - it takes practice! Solve lots of problems like advent of code, try some leetcode or Project Euler. For a particular problem, try lots of different approaches - think outside the box. Don't be afraid to take a hint to push you along. Lastly, it really helps to step away from the problem, think of something else for a while, then come back to it when there's not much pressure. I came up with my algorithm while on a long cycle! Good luck!!

@@mikelikesrobots Thanks so much Mike. New sub!!!

Welcome to the channel!

This is great! This taught me what unzip and abs_diff are, which I don't think I've seen before. I also learned that sort_unstable is faster than sort except in special corner cases - I'll use this from now on. Thank you!

I wondered whether to do this! I'll start from day 10, as I've already recorded day 9. Appreciate the feedback.

I'm not sure how that would work. Using a hashmap for rules allows an O(1) check for any rules that apply to a page number. Can you explain your suggestion more?

@@mikelikesrobots if you store the page number list as a hash map like "75->0, 47->1, 61->2, 53->3, 29->4", then to check if it's legal you have to iterate over all the rules (Vec<(pnum1,pnum2)>), and check if they are valid (both page numbers are in the hash map, and page 1 is before page 2: hm[pnum1] < hm[pnum2]

@@mikelikesrobots if you store the page numbers not in the direct form (vec, ord -> pagenr), but as its inverse: HashMap<pagenr, ord> (for example, hm[75] = 0, hm[47]=1, etc), then you can store all the rules as a simple Vec<(pagenr1, pagenr2)>. To check a list of page numbers we just have to check for each rule if it's applicable (both pages are present in hash map as keys), and if the order is right : hm[pagenr1] < hm[pagenr2]

That's a great solution! Then you iterate over the rules once instead of the numbers. I'm unlikely to code it as there are more AoC challenges I'm likely to struggle with yet, but I'll bear this in mind if the opportunity comes up again!

I'd recommend using a larger system for AWS Greengrass, as it uses Java and needs a good amount of memory to run correctly. But, who knows what they're cooking up at AWS!

@@mikelikesrobots I got it to work with the RaspberryPi 5. Is it easy to hookup the ESP32 and STM32 to see All the values of every sensor on them? Also, what else can these things do besides send sensor data?

What are you using for your AWS Greengrass?

That's great, the RPi 5 should have a good amount of memory to support it. Hooking up the ESP32 and STM32 is more up to you - if you can write the software on the RPi to read the values, you can use Greengrass to send that sensor data. There's a bunch you can do with Greengrass in general. Sending data is one, receiving messages from the cloud is another; you can run ML or Lambdas at the edge, get secrets from Secrets Manager.. Not to mention that once you can send data into the cloud, there is a lot you can do with that data. Many of my videos on this channel are exactly about answering what you can do with your robot data and the cloud. You can also see more of a curated view in my Getting Started page on my blog: mikelikesrobots.github.io/docs/getting-started/intro

Currently, nothing, as I have no active developments. However, my next robot project I would use it for OTA deployment of software and command and control, at a minimum. I would then extend it to push logs and metrics into the cloud so I can track my robots with a dashboard :)

Glad you liked it! You're right about the brackets, and I included a small bit at the start of today's video to explain it (out later today) - thanks for pointing it out. Good luck in your AoC!

Thank you! I'll try to finish the whole advent of code - eventually..

Of course not! You can work wherever you prefer. There are advantages and disadvantages of every job. I will say that Amazon has a lot of excellent engineers and a good attitude towards mentoring, but these are also available elsewhere. Best of luck finding something that suits you!

Have you seen the pre-built Greengrass component here? aws.amazon.com/greengrass/ml/ Otherwise, happy to help - what would the ML component be doing?

I'm glad to hear it was helpful!

Definitely! In fact, if you take a look at this video, it has a thorough guide with explanation of how to set up an environment like that: ruclips.net/video/pozA2Fc9kj4/видео.html

@ thank you so much

I'm so happy to hear that! It's exactly what I was hoping for when I made the video, as I'd been through the same amount of reading through documentation online and struggling to understand. Let me know if you have any suggestions for follow-up videos!

Glad you liked the video! I may be able to help, depending on the task. You can email me at mikelikesrobots@outlook.com if you want to discuss further.

I must have got it from somewhere!

It's a good question. In short, yes, but it will be more challenging. Big tech companies generally pick the best candidates, and they will be looking for college students. However, with impressive work experience and projects you can talk about, it's definitely possible. In fact, most companies have initiatives to support training existing employees with no formal education into these roles!

That's great to hear! I hope the videos continue to be useful for you!

@@mikelikesrobots "Thank you! Your guidance is really helping me learn a lot!" 😊

It is possible to get into the field without a degree, but it will be more difficult! Bear in mind that you're competing with people that do have degrees. If you're willing to learn on the job based on what robotics companies need, you can do it. I stand by my advice in the video - a mentor is great, and practical experience will teach you so much, if you can get an internship or similar. Good luck!

@@mikelikesrobots Thank you so much for taking the time to reply. I’ve decided that I’m going to “speed-run” a software engineering degree at WGU. Seems virtually impossible without it, especially in this type of job market. I will also be applying all of your advice. Thank you!

That's very kind of you! Glad to have you, and I hope the videos are helpful.

There is! You can find the build here: ruclips.net/video/q2bKTyTTquU/видео.html Or the whole playlist (that I'll be updating over time): ruclips.net/p/PLBrq1OKRHMwUbbujTlmt1YGRzL9O0LfNJ&si=MsOihpSEo9bng2Hu

@@mikelikesrobots ah thanks! I've watched it now I think I'll be able to assemble it

I'm glad you liked it!

I'd love to! The idea I have is to make ROS work with it, and buy a camera module to stick on the front. I don't know if a lidar module will work, but I'll look into it.

@@mikelikesrobots Thanks so much! I await the next videos with great anticipation!

Glad you found it useful!

Glad you enjoyed it!

You're welcome!

Glad you enjoyed it! That's a great question - I've touched on those areas but I don't have an in-depth understanding. I'd say get a lot of experience with ML, as many models as you can build. I've heard it said that ML is just as much an art form as it is science. A great grasp of statistics is a must. I'd also say that I'd prefer a conventional algorithm to an ML one so that it's repeatable, but at least I would hope that ML could interface with conventional algorithms, so you could research how to include normal code with your ML algorithm - maybe multiple models with code layers in between. Good luck in an exciting career path!

That's great to hear! I've also struggled with the setup, including for this video, so hearing you've been able to get up and running because of it is excellent.

I'm glad it was helpful!

This is a great goal!

🦄

Glad you liked it!

Hope it helped!!

gg!

That's a great point! It ties in well with admitting your mistakes and improving from them. Pinning this so future viewers can benefit from your advice.

That's great! Sounds like you have a solid grounding for building robotics. Maybe it's time to build a robot for yourself!