In 2013, at the scientific exhibition in Moscow, I saw an interactive promotional robot. The idea seemed interesting to me, but I did not have the experience to build a copy. At that time I was finishing 4th year university and worked for myself in the field of security systems. I went to the radio circle in my childhood, I understood well why we need these or other radio elements, but there was no experience with programming microcontrollers and setting up unix-like systems. To gain experience in this area, I decided to collect a copy of this robot, improving some points. Of course, it would be foolish to start from such a project right away, and a 3d cube of 5x5x5 was assembled as a test, then the 3d cube was collected by a more serious resolution of 10x10x10 pixels:

Having accumulated more experience on simple projects, it was decided to embark on a long-standing idea, namely to collect a promotional robot. In the best traditions of the genre in the kitchen in the evening they discussed the functionality and laid down the basic principles of construction. If in principle it was clear with the platform and the technical implementation, then we had no idea how to make the body, head, and generally beautiful appearance.
At that time, the Rapberry PI mini-computer was selected as the main platform, and a homemade AtMega32 board was used to control the power elements. Since I was “friends” with video surveillance systems, it was decided to use an ip camera with MJPEG stream encoding as a robot's eyes to receive a stream to any device through a router.
')
A router with an external antenna serves to connect to a wi-fi point, which makes it easier to control the robot from any device with a browser. Also through it, we transmit the stream from the ip-camera, so as not to download raspberry. A 24-inch monitor on the body of the robot is designed to display a promotional material in full-hd resolution. Also in the head of the robot there are two “eyes” from a matrix of bright LEDs, to display emotions with a resolution of 10x10 pixels each:

Now the robot is almost ready, and very soon we will send it to a beta test for a promotion in the shopping center of our city (we are in Rostov-on-Don). In the framework of several articles I will tell you what difficulties we encountered in the process of building a project. In the first article I will tell how we make the body. Looking ahead to say that this was one of the most difficult tasks.
For the test (we naively thought that we would immediately make a super-design) we came up with a simple tapered hull design, which no one in Rostov was trying to make according to the drawings. A lot of information gathered in the automotive forums, in the reports on the manufacture of non-standard bumpers. It was decided to make a model for the subsequent manufacture of copies of fiberglass from it.
For the blank, 4 walls of the cone and radial parts were cut to maintain the evenness of the circles:

The assembled skeleton was neatly filled with foam, and then the main mistake was covered. The construction was not very smooth, but I was inspired with confidence that everything can be corrected in the following stages:

The resulting disc was decided to strengthen the fiberglass, for this was bought epoxy resin and hardener:

Then, in order to align the structure, a lot of putty was left, in some places 2 cm each, and in general I was not satisfied with the result. The hull weight was about 18 kg, which was no good:

It was decided to use this body as a blank, and using it, make a light body. The pig was plastered with layers of papier-mâché, to maintain the stability of the proportions:

And then the glass fiber was already applied on top, only as a resin it was decided to use polyester resin, and not epoxy. Buying gloves and respirators, we got down to business, and got a light construction, ready for alignment with a minimal layer of putty:

Then leveled with putty:

Then they disassembled the monitor and made a box for the monitor from aluminum plates, made holes for the screws that hold the monitor and attached this whole structure to the case. Plywood was cut out on the sides, which bonded the case and put it all into one:

The alignment work was already debugged and the body was aligned to the “not ashamed to show” state:

Pre-stained for tests:

As a head for the dough after papier-mâché technology, we didn’t think of anything better than how to strengthen an ordinary inflatable ball, especially as we liked its shape. The result was pretty good:


The question now is how to make a translucent glass of such a shape that repeats a ball, there are solutions and they are tried. Probably it will be thermal vacuum molding of plexiglass:

In the next article I will tell you how the technical part was done, what problems we faced and possible optimization options.