GPS

[Michele Perrone]

Posted on July 18, 2017 at 15:40

Hi, before buying I preferred to ask advice to you. I am pursuing a project that has the ambitious goal of bringing analytical accounting to farms. To do this I have to automate some processes. The process I now describe to you. I have to locate the perimeter of a geographic area. That is what I must identify who comes in and who comes out of this area. So I need a device on which to load the perimeter of a certain area. Also, this device every time it enters and exits that area, I must send a signal with information. This device must be equipped with an internal memory to store this information inside it. Which device do I recommend to buy at ST to start testing? In addition, this device has tutorials or documentation in general to allow me to program it. Thank you for your time. I await your news.

[valentin]

Posted on July 18, 2017 at 22:31

Hi, maybe you focus on the software side of that project first? Once you've established the correct algorithms you can then tailor-fit your hardware to that.

I would approach that using my phone as a sensor -> I'm sure there are some apps that allow exporting raw GPS sensor data to the internal sd card. Maybe export IMU data as well (accel, gyro, magnet). Then you can use an unscented kalman filter to get a high-resolution 6-dof position of whatever you're tracking. Once you have that, you can then setup a method to define the area of interest and according onEnter and onExit triggers. I would do all my prototyping in Matlab. Once it all works in there with your phone data, you can convert it to an actual embedded system. Does that make sense?

[Tesla DeLorean]

Posted on July 18, 2017 at 23:25

So basically Geo-Fencing.

The level of accuracy you want will determine the type of GPS required, and what infrastructure will be needed to support that. Regular GPS typically provides accuracy with a radius in the 3-5m range. PPP might get you to a metre, and RTK to centimetres.

The receivers will tell you the location, the line vectors and geo-fencing is on you.

Areas bounded by circles are simpler to implement.

[Michele Perrone]

Posted on July 19, 2017 at 16:05

I thank Valentin and Clive One for answering me. I live in Italy and therefore apologize for my english not good.

I have tried to find information from you because I have done economics studies. I in computer science find myself with the first weapons.These economic studies have led me to draft the form, settings and flows of this software structure.For more than three years I've been working on this softwrae.I am motivated by the fact that there is a great need for analytical accounting in companies.Unfortunately I do not have much knowledge of computer science and so I have to learn many things.I would be very happy if you could drive me into this learning process.I want to understand what the various steps I need to do to achieve the final output.I want to understand which documents I must read.I want to understand what I need to buy.The project is based on analytical accounting on farms.The question is very wide, but the first problem I have to solve is GPS tracking.I'm a banal example.The example concerns an agricultural company.A farm has an employee who goes to his production site. The person entered this land and carries out one or more operations.The data I need to capture is as follows:

The identity of the person who entered the field;

The hours left in the field;The operation or the operations performed by the person.

As for the operation. I must have previously included in the program the operations that can be carried out on that ground. Each operation has distinctive features.

When a person leaves the field, the program must trace back to the type of operation he has performed.Fundamental is to understand what kind of operation the person has played. In fact, the form I have drafted manages the costs according to the nature of the transaction. So depending on whether the transaction is ordinary or extraordinary it will undergo its economic management.

I'll send you my mail

mailto:info@tralcio.com

I have not posted the site for work commitments but I will post it shortly. I am also available for corporate equity remuneration if you would also like to help me materially.

To do what I told you about what to start? If I have to buy qaulcosa, what should I buy?

Thank you

[Tesla DeLorean]

Posted on July 19, 2017 at 19:41

So realistically a fairly complicated hw/sw project where you'll need to hire several engineers with appropriate experience.

If you want to determine if GPS provides suitable data, you could get a USB G-Mouse type device, and walk your environment, or proxy for it. Decide if the positional information provides the precision, accuracy and repeatability you need.

http://www.navilock.de/produkte/G_62456/merkmale.html

 

http://www.navilock.de/produkte/G_62524/merkmale.html

 

Then you'll need to generate your own mapping data bounding areas of your environment, and process that with respect to the real-time positional data provided to you by the GPS receiver.

[valentin]

Posted on July 19, 2017 at 22:45

Hmm, the project looks very promising indeed. But it seems like you're doing this all by yourself? If you have a look around at successful business startup stories, you will find that they all teamed up with someone eventually - but mostly at a very early stage. That's what I would recommend to you, too. Find a friend, buddy, partner who is more experienced in the technical side so you two can do what each one does best. You do economics and management and he takes care of technology.

It IS possible for you to do it all on your own, yes. But it will take you WAY longer and therefore reduce chances for success a lot.

If you want to give it a shot anyway, just get yourself the cheapest and easiest to use GPS receiver and start using it to generate some raw data. For example, walk around a nearby big lawn and then in and out of that lawn and try to simulate a working process. Then have a look at that data and visualize it (plot). You might find that you don't need very high accuracy because of the large scale of things.

My strategy usually is: start as simple as possible and then work my way up from there.