User Case Study: Maximizing energy regeneration with MapleSim makes improvements to assistive devices possible

Share

Most research of human, animal and robotic motion has been centered on improving everyday activities such as walking, running, turning, starting, accelerating and decelerating. However, not much attention has been paid to assisting movements such as standing up and sitting down - movements that seem relatively simple but become increasingly difficult with age and reduced health.

Over the last few years researchers have begun to make advances in producing more practical and streamlined devices to aid these movements. A wider interest has also recently developed in humanoid robot communities to build upon the existing assistive mechanisms by improving their energy efficiency.

James Andrew Smith, currently a faculty member in York University's Lassonde School of Engineering, and a team of researchers at Ryerson University, have done advanced research on autonomous battery operations in humanoid robots and electrical assistive devices using MapleSim, the system-level modeling and simulation tool from Maplesoft. Smith’s group undertook the task of determining at what point in the transitions between sitting and standing can energy be regenerated in an orthosis or prosthesis, much like how a hybrid vehicle regenerates energy during braking by drawing it from the motor for re-use in the vehicle’s operation.

Determining which of the three joints - the ankle, knee, or hip - is able to regenerate the most energy in the sitting down and standing up movements would prove to be a practical and significant consideration for rehabilitation engineering design. The conclusion could lead to more efficient locomotive devices for those who suffer from diseases or disabilities affecting the muscles around these joints.

Turning to the MapleSim battery modeling library saved our team considerable time and effort.

— Dr. James Smith, York University

In order to successfully determine at which point regenerative power is at its peak, Smith’s group applied biomechanical data from actual human trials to a robotic model created in MapleSim. The robotic model mimics human movements when transitioning between sitting and standing positions. To investigate regeneration at the ankle, knee and hip specifically, MapleSim models of an electromechanical subsystem actuator, a DC-to-DC bridge converter, and a battery were placed at each localized joint.

To produce an optimal rehabilitative device that is feasible for everyday use, the system would have to run on its own power supply. Smith’s research group used a regenerative braking circuit with an actuator, battery, and H-bridge to analyze energy consumption and regeneration during the various phases of the sitting down and standing up movements. The circuit harnessed the power linked with the counter electromotive force voltage of the actuator to charge the battery. Normally, the battery would provide positive power to an electric motor, but in the regenerative braking circuit, the motor acts as a generator so negative power can allow energy to flow back into the battery.

“Our team needed to be able to model the battery’s complex chemical reactions, a feature that has traditionally been difficult to find in many engineering software programs,” Smith said. MapleSim’s battery model accounted for the electrochemical processes and thermodynamic behavior of the NiMH battery and described these equations as a set of equivalent electrical components interconnected with one another. Smith noted, “Turning to the MapleSim battery modeling library saved our team considerable time and effort, as we did not need to model the battery from scratch; we could begin with an already-advanced model and simply edit it to fit our project.” Smith’s team also used the MapleSim H-Bridge DC-to-DC converter, a model available on the Maplesoft website, to handle the electrical power transfer between the NiMH battery and the actuator.

Smith’s research group developed two simulation models using MapleSim. First, a simplified model was created with the foot firmly attached to the ground. This permitted the researchers to create an efficient, low complexity model-based motion controller. This controller was then applied to a more complex and realistic human model with feet that could lift off the ground. Data from human trials was used to provide the desired trajectories for the simulations in a multi-domain model in MapleSim. The design was constrained to a 1/10 scale in order to implement the model in a small robot using a Dynamixel RX-28 actuator. “Using MapleSim’s multi-domain environment, we were able to accurately simulate the necessary motions in order to properly analyze battery autonomy,” explained Smith.

Kinematic robot model with foot, three actuated joints and two points of contact between foot and ground.

To determine the point at which the most regeneration of energy occurred, the researchers applied the Dynamixel RX-28 actuator to the scale humanoid robot during the biomechanically-accurate movements. During each transition between sitting and standing, the state-of-charge was plotted over time for the hip, knee, and ankle as an indicator of the battery capacity at each stage.

Simulation of the sitting down and standing up movements began with the model sitting on a virtual chair. As the simulated robot rose from the chair, the state of charge decreased over time as power was drawn from the battery. Once the body was above the foot, the hip began to brake and regenerate energy. This occurred also in the sit-down motion when the body maintained its center of mass over the foot. Very little regeneration occurred overall for the ankle and knee subsystems compared to the hip.

Smith’s group also demonstrated that regeneration is more prominent during the sitting down phase than during the standing up phase.

 

Smith’s group’s findings have a meaningful application for prostheses and orthoses design. Determining the most efficient battery autonomy means the operation time of these devices can be extended, and smaller, lighter batteries can be used, reducing their bulkiness. Ultimately, a more efficient device can reduce the joint loads during standing-to-sitting for users - a critical consideration for people suffering from joint diseases.