This article presents a novel control framework for assisting human arm movements with a robotic device. Unlike existing collaborative and wearable robots that predominantly operate in task space or joint space, our framework focuses on controlling variables related to the muscular space, which refers to the activation and force generation of individual muscles. Translating these variables to the robot's actuation space is challenging, which hinders the development of human-centered tasks and limits the transparency of robotic assistance for the central nervous system. To address these issues, we consider the relationship between the muscular space and the robot's task space, using the kinematics and dynamics of the human limb obtained from a calibrated musculoskeletal model based on Hill's muscle model. Our framework includes two muscle-targeted methods: assistance and augmentation. The assistance method allows exercise or rehabilitation of specific muscles, while the augmentation method achieves isotropic manipulability, enabling the user to exert forces uniformly in all directions using the novel concept of a force envelope. The experiments were conducted using a haptic robot in admittance mode with a viscous environment simulating a load. The results demonstrate significant improvement in the prediction capabilities of the calibrated model compared with the baseline. Muscle-targeted assistance induces substantial changes in targeted muscular efforts. In addition, we achieve isotropic manipulability in a more precise way as compared with previous methods. In summary, our proposed control framework proves effective in assisting human arm movements with a robotic device for rehabilitation and power augmentation in human-centered applications.

This article addresses a new way of generating compliant trajectories for control using movement primitives to allow physical human–robot interaction where parallel robots (PRs) are involved. PRs are suitable for tasks requiring precision and performance because of their robust behavior. However, two fundamental issues must be resolved to ensure safe operation: first, the force exerted on the human must be controlled and limited, and second, Type II singularities should be avoided to keep complete control of the robot. We offer a unified solution under the dynamic movement primitives (DMP) framework to tackle both tasks simultaneously. DMPs are used to get an abstract representation for movement generation and are involved in broad areas, such as imitation learning and movement recognition. For force control, we design an admittance controller intrinsically defined within the DMP structure, and subsequently, the Type II singularity evasion layer is added to the system. Both the admittance controller and the evader exploit the dynamic behavior of the DMP and its properties related to invariance and temporal coupling, and the whole system is deployed in a real PR meant for knee rehabilitation. The results show the capability of the system to perform safe rehabilitation exercises.

As it is well known, robots are complex devices capable of performing very fast and precise movements. Developing robots and their controllers is a very challenge task due to the different technologies that are needed to be dealt with (access to peripherals, real-time operating systems, communications ...) and the amounts of programming work that is needed. Moreover, this work needs to be repeated in most cases when a new robot is developed. A new way is needed for developing the code implemented for controlling these robots that allows to reuse it in a safe way. In this paper, the Orocos (Open Robot Control Software) middleware has been chosen for the controller development. Orocos is a real-time middleware, focused on control systems, especially those related to robotics and automation. Its greatest advantage over the other solutions available is the capability to provide an off-the-shelf hard real-time operation. This is essential in most of the robotics applications, making this middleware in a very suitable piece of software. Since Orocos is a component-based middleware, several CBSD (component-based software development) techniques have been used to design and implement the control system. Thus, using a modular control structure, a number of advantages can be achieved such as code reusability, execution of the modules in a distributed way, ability to load or instantiate a module several times, easy following of flow execution and fewer programming errors. In addition, these components are configurable and reconfigurable in runtime.