Autonomous Hexapod Design with Gait Free Planning for Rough Terrain

Project Title

Autonomous Hexapod Design with Gait Free Planning for Rough Terrain

Project Area of Specialization RoboticsProject Summary

Walking robots have been researched during the last four decades in an effort to replicate the natural skills displayed by various animals and insects. Nonetheless, mobility and stability in these applications are difficult particularly on uneven terrains. Wheel-based robots are not stable in rough terrains. Robots like Boston Dynamics’ 4-legged robots are expensive and they have a hydraulic system.

Nonetheless, the transition from wheel-based to leg-based robots has proven to be more efficient and stable on rugged terrains. Still, robots with four or fewer legs have more issues maintaining their center of gravity to avoid tipping over.

However, Hexapods can outperform both in terms of speed and stability.

Hexapod means six-footed. It is a six-legged biomimetic robot that mimics insect movement. It is a fully mobile robot that can walk in many directions with self-balancing capability.

Our robot will have two modes. One is App control mode and the other is Autonomous Mode. In App control mode, sensors will continuously send the surrounding information to the user App. Every movement will be controlled by the app. In Autonomous mode, The hexapod decides its gait according to the rough terrain and obstacle. This makes it gait free Hexapod.

The robot’s behavior (motion and manipulation) is designed according to the tasks that the robot can achieve. These tasks are selected according to the size of the motion space and the importance of the task. The design method relies on dividing the tasks (according to their priority) into different levels. An efficient method that prevents conflict between the tasks will provide an opportunity to perform multiple tasks simultaneously. Indicating that the overall body stability is the highest priority that should be satisfied first. While there are many methods suggested for the stability criteria for legged robots, most of them are based on the same fundamental concept, which is to keep the Center of Gravity (CoG) of the robot inside the supporting legs. For the purpose of using a six-legged robot in for example minefield, a statically stable method would be appropriate as the robot has to walk very slowly. Therefore, accelerating the process depends on performing multiple actions during the walking task.

Project Objectives

..

The aim of this project is to design an autonomous hexapod robot and its control system, with the capability to avoid obstacles. Furthermore, to distinguish the type of terrain and accordingly decide on the gait type and legs configuration. The challenge is how to control the interacting systems of the body movement, body attitude adjustment, manipulator sweep, and leg foothold. Integrating all these characteristics in a legged robot will increase its overall performance.

The traditional methods are very dangerous and difficult task; any site of previous hostilities is extremely hazardous with many terrains such as bomb craters and many uncertainties as to the whereabouts of ordnance. The problem of landmine would be partially solved if mines could be reliably detected, identified and accurately marked. Although this is a considerable challenge for robotic researchers, legged robots might be an effective and efficient means of detecting, and marking mines while guaranteeing the safety of people engaged in the clearing task. Although there have been many attempts to use six legged robots in dangerous areas; however, these attempts are limited to conventional remote control models. The new trend in control strategies is based on tasks that can be achieved by an autonomous robot. The overall performance of the robot can be improved if all its motion space is used optimally to perform more than one task at the same time.

In order to achieve the above mentioned aim, the objectives of this thesis are summarised in the following points:

• Defining the robot tasks using operational space techniques.

• Generating continuous walking in a six-legged robot by defining the motion reference at the robot’s CoG. The reference path includes the desired direction and orientation of the robot’s body. The trajectories for each leg is generated and constrained by the robot’s body path.

• Tracking a certain straight-line path defined over the ground.

• Reducing the dynamic effect that is generated by the manipulator on the robot’s body by decomposing the contact forces at each leg to normal forces.

• Evaluating the effectiveness of the algorithm using different motion and scanning scenarios.

• Making this robot capable of being used in industry primarily focused on use cases where wheel based robots are not the solution.

• Increased Stability controls for robots to have flexibility over the rough terrain.

Project Implementation Method

..

Block Diagram:

Flow Chart:

The traditional hexapod method of fixed gait was not suitable for moving the robot on rough terrains, so the design and implementation of newly made gait system was necessary to make the hexapod capable enough to tackle with the problems presented with like having capability to kater the stability issues. The method which has been researched in the recent few years is using a tri gate mechanism which makes the robot move 3 legs opposite to each other first and then the other three legs in the next pass. This was then transferred to using a free gait mechanism which decides the gait of robot motion depending on the surface like if the surface is plain or it’s not plain instead it’s rough. So the sensors will eventually sense the surface and give information to the robot microcontroller which eventually decides the gait for the robot about how much a leg is needed to move with certain angles to each servo attached to a single leg. In this way, each leg moves independently of all other legs thus making it easy for the robot to have flexibility of moving in rough surfaces along with a strong stability issue resolved. The another positive aspect of this project is to implement a machine learning model using raspberry pi which will do image processing to make decisions about how the surface ahead of the robot is and whether robot should move forward surpassing the rough terrain or it’s much more uneven surface which is difficult for hexapod to surpass. In the second case, the robot will just change its direction  thus avoiding that path ahead and taking another route to move forward. This all is the case when a robot moves in Autonomous Mode. In the case of App Control Mode, The robot can be redirected to any direction using app control by the user and thus the robot is responsible for its stability only and not the path to choose.

In order to perform a walking task in a legged robot, the robot’s legs must initiate proper contacts with the ground. Part of the available motion space is utilised to satisfy some constraints that arise from the interaction with the environment. The adopted method is to design a simulation model using Matlab and modelling all physical parameters.

The overall mobility of the robot is achieved by defining a set of control points in the task space, such as the robot’s CoG, legs, feet.. The associated constraints are formulated as linear equality constraints (e.g., zero velocity at the contact point) and linear inequality constraints (e.g., joint limit and obstacle avoidance). The approach is based on projecting the tasks with low priority level in the null-space of the higher priority one. Typically, the entire balance is affected when the robot encounters an obstacle inside its path during the walking task; to overcome this problem, the inverse kinematics method is adopted to get a compliant response.

Benefits of the Project

While our primary goal with respect to the hexapod design is to implement an algorithm that works in tandem with the microcontroller to ensure accurate synchronization, the future applications of such a design are much vaster.

•Our design can be used for automation purposes as well since it is essentially based on the synchronization of inputs based on situations.

•Our hexapod design can be used for space exploration as well since its independent appendages allow it to be more versatile in its movement capabilities.

•The sensory system can be used for geological research sites where conventional wheel-based systems cannot function properly.

•Can be used for monitoring agricultural fields for pest control.

•Can be used for search and rescue operations.

•Can be used to survey volcanic regions which are inaccessible by conventional robots.

•Can also be used to survey narrow pipes and air ducts that are too compact for humans.

•Just like NASA is using autonomous robots to explore caves on Mars. Our robot would be a suitable replacement.

Technical Details of Final Deliverable

..

Robots face many uncertainties during achieving its work, controlling legged robots using traditional methods, such as position control, is not applicable. Dealing with the environment represents one of the most factors that generate these uncertainties. The six legged robot has at least three contact points with the ground (case of tripod gait). These contacts impose more constraints on the robot’s motion and need to be satisfied before performing any task. Therefore, eliminating the effect of the contact forces using decomposition methods is insufficient in case of the six legged robot due to there being three or more contacts with the ground. Ultimately, having achieved the continuity motion in multi-legged robots in uneven ground and in presence of external forces will open up the use of these robots in applications that require quasi-static or dynamic balance rather than static balance.

Some of the final Deliverables of the project are as follows:

• Performing the matlab simulation of hexapod robot to test the gait design before its actual hardware based implementation.

• Development of hardware based solution with a unique and compact design or hexapod skeleton having support to attach all of the components deliberately.

• Developing the code for the gait design to make hexapods run and work in the same way as illustrated in matlab simulation based results.

• Combining all the hardware components and complete and working Hexapod robot with functionalities as described in the objectives of this project.

• Compiling a project report demonstrating the complete working principles of the robot along with technical prospects of the gait planning and obstacle avoidance mechanism using image processing.

Final Deliverable of the Project HW/SW integrated systemCore Industry AgricultureOther Industries IT , Medical , Others Core Technology RoboticsOther Technologies Artificial Intelligence(AI), Internet of Things (IoT), OthersSustainable Development Goals Affordable and Clean Energy, Decent Work and Economic GrowthRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	62930
Hexapod skeleton kit	Equipment	1	14000	14000
Servo motors	Equipment	20	550	11000
stm32	Equipment	1	6000	6000
LM2576T 5.0	Miscellaneous	2	50	100
Cell holder	Miscellaneous	3	50	150
XL 4015 BUCK DC to DC	Equipment	1	200	200
BMS 3S 10A ORG	Equipment	2	180	360
Wires	Miscellaneous	1	90	90
3.7 18650 capacity battery	Equipment	9	170	1530
raspberry pi 4	Equipment	1	28000	28000
Servo driver	Equipment	2	750	1500