2.4 Tailored Behavior State Machine of OpenPlanner for
Greenhouse
A behavior state machine is utilized to guide the robot’s specific
reactions under varying conditions. This finite state machine is a
computational model often employed in applications requiring definitive
transitions between different states (Conner and Willis, 2017). For
example, whether to approach or avoid a plant upon encounter is
dependent on the specific task being executed.
A behavior state transition table designed for greenhouse environments
is shown in Table 1, which can be adjusted based on actual work
requirements.
However, due to numerous disturbances in agricultural environments, the
robot reacts to each minor change. This reaction triggers frequent state
transitions, consequently causing system instability. To mitigate this
problem, a hysteresis execution strategy is introduced in this paper.
The timer is initiated upon the detection of a potential state
transition. The state transition only ensues if the conditions are
sustained for a predetermined period, set at 0.5 seconds in this study.
Should the conditions not persist for this duration, the timer is reset
to prevent the state transition.
Table 1 . Behavior State Machine Designed for Greenhouse
Environment