Recently development of the improved algorithm of the computer sight allowing the robot to calculate own movings (a visual odometriya) began. The algorithm compares among themselves the images received from chambers of the robot during the different moments of time. Further it finds earlier seen characteristic parts of the previous image on the new image and tries to define, on how many they were displaced. Then, on the basis of these data, calculation of moving of the robot is made. At present completion of elements of the algorithm which is responsible for allocation of characteristic parts of the image and calculation of shift of the robot is conducted. As the module allowing the robot was improved to analyze information from the laser scanner, for the purpose of detection of obstacles along the line. By means of the received data the robot models routes of a detour of obstacles, in passing updating information on these obstacles. The Supports module which works by a principle was improved - ”saw object and I follow it”. Now the prototype remembers object better and follows it quicker, than earlier.