Planning reaching and grasping movements: the problem of obstacle avoidance

Area of Science:

• Robotics
• Biomechanics
• Cognitive Science

Background:

• Human movement planning involves complex processes for reaching, grasping, and obstacle avoidance.
• Existing models often simplify obstacle interaction, limiting their predictive power.

Purpose of the Study:

• To present and validate a computational model of human movement planning for reaching and grasping tasks.
• To investigate the role of instance retrieval and generation in multi-stage movement planning.
• To compare model predictions with observed human hand kinematics during obstacle avoidance.

Main Methods:

• A computational model simulating one- or two-stage planning processes was developed.
• Instance retrieval and generation principles were applied within each planning stage.
• Model performance was evaluated against empirical data from subjects performing reaching tasks with and without obstacles.

Main Results:

• The model successfully predicts trajectories for direct reaching and obstacle avoidance.
• A two-stage planning process, involving a 'via posture', effectively resolves collisions.
• Model kinematics align with observed human movement patterns.

Conclusions:

• The proposed model offers a robust framework for understanding human movement planning, particularly in obstacle navigation.
• The model's success in replicating observed kinematics supports its validity.
• This model has potential applications in understanding and treating motor disorders like spastic hemiparesis.