The paper discusses the shape control problem related to a class of hyper-redundant robot arms with continuum elements, i.e. tentacle arms. A spatial weighted technique for sensor measurements is used in order to facilitate the parameter estimation. The paper focuses on the shape control by using the curvature gradient a constant parameter along the segment arm. The conditions that ensure a constant curvature gradient for a class of tentacle arms characterized by elastic backbone are determined. A sensor network distributed along the robot arm is used for the shape control. The main parameters of the arm shape, curvature and curvature gradient or "shape" Jacobian for the control problem are estimated. Two measuring systems are used: a) a distributed angle sensor network and b) a curvature sensor placed at the end of the arm segment. The stability analysis and the resulting controllers are obtained using the concept of boundary geometric control and the weighted state control methods. The shape control algorithms for dynamic models with uncertain components are proposed. Numerical simulations and experimental results illustrate the effectiveness of the above mentioned algorithms.