**2. Related work**

Climbing and walking robots are common in industry applications, like as, maintenance activities of nuclear plants, oil refineries, bridges, high voltage towers, medical fields, endoscope devices, and surgical instruments in general. The applications mentioned above are practiced with some problems to access and with hazardous environmental conditions.

To accomplish these tasks, robots need some important features, as reliable or robust, to be able to move over 3D structures, including walls, ceilings, pipes or cylindrical structures (Reinoso et. al, 2001), and they also need to be adaptable into different terrains.

Parallel robots have good performance, and they are perfect to manage manipulation tasks with short manipulation cycles, high speeds and accelerations. These characteristics are very difficult to obtain in serial robots. However, the proposal system combines advantages of serial and parallel robots. For example, thanks to the linear actuators, a parallel robot has a high ratio of payload and deadweight, so using linear actuators is totally justified.

Another desirable characteristics of these robots are that they need powerful torque in the actuators, mainly if they use serial legs for climbing. Although our system uses serial legs for climbing, the combination with parallel modules does that the torque of the actuators is not needed to be very high.

Higher speed is a desirable characteristic, but it is reduced when using legs for climbing, however, in our system, velocity is not a problem because of the parallel modules that are used, similar to Stewart-Gough platforms.

Most common problem in the walking and climbing robots is how to negotiate the boundary of two plain surfaces such as convex or concave corners in 3D. In this paper, we propose a robot model that solves this problem in a right way, using hybrid legs.

The main purpose of this project was to design a very simple robot similar to the Stewart-Gough platform, but combining serial robot abilities that allowed this to do climbing tasks. Probably the most important task to solve in this robot is the control system. To solve this, it is proposed a binary actuation. The binary degrees of freedom in this system are quasiexponential compared with serial robots, but the approach with a serial robot is increased too. As opposed, our system does not require feedback of any sensor, among other features. Some researchers have studied theses kind of binary actuation robots, but none of them has used hybrid serial-parallel technology.

Similar to parallel robots (Aracil et al., 2006a) this robot can climb the exterior and interior of tubes or metallic structures. According to the kind of structure, the end-effector and the base could carry magnetic foots or suction pads (Kim et al., 2005).

In this chapter, we do not dedicate attention to the fixation system of the end-effector and base of the robot, although a suction pad or magnetic feet are recommended. We are going to concentrate on the kinematics analysis and the postures it can achieve.
