Frequently Asked Questions (FAQs)

Our primary technical risks include ensuring the durability and resilience of our robots in harsh underwater conditions, refining our AI algorithms for optimal target identification and resource recovery, and maintaining reliable communication systems between our robots and surface operations. We are actively working on mitigating these risks through rigorous testing, continuous improvement, and strategic partnerships with leading technology providers.

A number of companies have conducted dredging tractor and riser system deep sea trials; however, the ship-to-ship transfer of the nodules has not been tested. The ship-to-ship transfer of the nodules in the ocean is complex and will likely require very specialized transport ships, including dynamic positioning (DP) and dewatering technology. This article includes more details on this topic.

Our technology will be ready for commercial mining operations by 2027. As of mid-2024, we are in the final stages of testing and validation, ensuring our systems meet all regulatory and environmental standards. We also engage with potential customers to align our development with market needs and expectations. We are currently designing Eureka III, the full-size production system.

We expect Impossible Metals’ approach to be the lowest cost method for deep sea mining. A fleet of robots has three primary economic benefits compared to dredge and riser-based systems.

The first is the improved economics for a fully operating system. This is achieved through reduced capital expenses (CapEx) by not requiring a dedicated surface production vessel with dynamic positioning for supporting equipment like a riser system. With the Impossible Metals approach, the transport ships pull the robots from the water without needing dynamic positioning surface vessels or ship-to-ship ore transfer. We need one ship instead of two.

The second benefit is the ability to scale the system with the incremental addition of CapEx. A small-scale operation can become operational with a relatively modest initial capital investment. As additional capital is invested, the fleet of robots and, thus, the material throughput can be scaled.

The third economic benefit is the lack of single points of failure. While there are increased points of failure with the fleet of robots, there are no single points of failure, ensuring that the selective harvesting architecture remains operational through these failures.

Reliability is a cornerstone of our technology development. Our robots are designed with multiple fail-safes, redundancy systems, and real-time monitoring capabilities to ensure consistent performance in various underwater conditions. We conduct extensive testing under simulated and real-world conditions to validate the reliability and durability of our systems.

We address the complexity of our technology through a modular design approach, enabling easy maintenance and upgrades. Our interdisciplinary team of experts continuously collaborates to integrate cutting-edge robotics, AI, and marine engineering advancements.

A mission plan is uploaded to each robot before deployment. The robot is fully autonomous and completes the mission plan based on sensors and programming. The robot’s progress is monitored through an acoustic modem (USBL) that tracks the robot’s position and provides low bandwidth status information. The fleet management software on the surface will automatically send essential speed and course adjustment information to the robots in the water to support synchronization of the robotic vehicle launch and recovery. Additionally, the operator on the surface vessel can provide override commands when required.

To allow many robots to operate in parallel, we need to eliminate the need to manage a tether (cable). Each robot is lithium-ion battery-powered.

The current end effector and internal nodule conveyor can accommodate a wide range of nodule sizes from 2 cm to 10 cm in diameter. Refinement to the end effector design and conveyor for the production system will be customized to the nodule distribution in the field of operation. Our concept economic model assumes an average nodule collection size of 7.5 cm.

The underwater robot incorporates several standard environmental awareness sensors, such as DVL (Doppler Velocity Log), pressure sensors, USBL (ultra-short baseline), and Stereo Cameras, which are fed into the INS (inertial navigation system), where the information is fused to develop a state estimation for the AUV. As multiple AUVs begin to operate concurrently, a dynamic LBL (Long BaseLine) positioning approach will also be employed.

More information about subsea positioning

The state estimation accuracy for the AUV will be highest at the seafloor, where it is in proximity to other AUVs that have maintained a high accuracy, and near the surface, where it is close to the surface vessel-based USBL. When at the surface, the position will be based almost exclusively on USBL. When at the seafloor, the dynamic LBL combined with DVL and vision tracking of the seafloor will dominate. The stereo cameras on the front of the AUV map out the locations of any nodules and any macro life for the arm control algorithms to determine how to handle them.

As the underwater robot travels through the mission, it adjusts the operating mode from having a desired depth while it travels down through the water, and then, as the acoustic sensors and optical sensors begin to observe the seafloor, the vehicle autonomously transitions to altitude control. Altitude control will be active while collecting nodules. The vertical thrusters adjust the force they exert to maintain the set altitude for the AUV over the seafloor. As nodules are collected, the buoyancy engines on board the AUV pump water out of the tanks to maintain the same small desired downward force from the vertical thrusters.

The seafloor’s typography, where we will collect nodules, gradually changes, and the vehicle travels relatively slowly. A forward-looking sonar provides feedback to the system for topography changes that are too much for the vertical thrusters to manage.

Each underwater robot will be provided with a mission as part of pre-launch communications. Through the ship-based USBL and seafloor dynamic LBL positioning, the robot will move to the desired target starting location. At the start and end of each path, the robot will perform a small visual survey. When arriving at the start of a run, the robot will survey the area briefly to ensure that the positional alignment of the new path is correctly positioned relative to the start or end of the robot from past paths. It will use the distinctive pattern of nodules in the survey as a fingerprint to understand any centimeter-level misalignments that the navigation system is experiencing relative to the navigation system from other robot trips to the same location.

In our economics, we model a 25-year effective life for the underwater robots. Underwater vehicles, when maintained, can last well beyond 25 years. The JASON ROV launched in 1988 is a good example of this. We will spend about 10% of the total CAPEX each year on the maintenance of our equipment. This high degree of maintenance ensures that continued operation is similar to an airplane’s.

The main factors that influence underwater robots in general are maintenance management for:

• Corrosion: Low-corrosion materials and sacrificial anodes are used and replaced as needed.
• Biofouling: The ongoing maintenance will include a maintenance schedule for clearing acoustic and optical surfaces to prevent biofouling build-up.
• Moving component wear and tear: The primary cost factor for maintenance is rebuilding components due to wear and tear or in the event of failure, and we’ve incorporated it into the economic model. (Learn more about analysis in this blog post.)
• Electronics failure: In the event of a complete loss of power or control, the robot will always remain positively buoyant and float to the surface. 

Since we are operating a large fleet of underwater robots, we can collect performance and required maintenance data on a statistically significant scale and employ maintenance insights based on this information.

Eureka III delivers four metric tons every 3 hours. Eureka IV delivers 12 metric tons every 4 hours.

Before deployment, the mission is wirelessly uploaded via WiFi to the underwater robot while on deck. During operation, deviations from this initial plan can be uploaded to the robot via the acoustic link. The plan that is initially uploaded and the following deviations are provided by the master software, which is constantly developing updated operational optimization.

Today, we use a diesel generator to recharge the batteries. In our economic model, we capture the carbon from the diesel generator or pay for carbon offsetting. In the long term, we will look for a renewable power source, such as waves, wind, etc.

The response of the underwater robot depends on the nature of the emergency or failure. Awareness of emergencies and failures external to the individual robot will be communicated acoustically through the USBL, and the updated desired behavior of the robot will react accordingly. For example, suppose the surface vessel becomes unavailable due to incoming weather conditions or an onboard emergency such as a fire. In that case, the communication signal will be sent to the robot to change its course or hold in place until the weather has passed or operations can be resumed in extreme cases with the second ship in the operations. 

For small internal failures to the robot, such as loss of connection to certain computers or components, an isolated power reset will occur to re-establish the connection. If the error persists, the mission will abort. Other small failures, for example, could be a break to an arm, and in these cases, the robot is expected to complete the mission and have the damage repaired at the surface.

Each vehicle has a high degree of redundancy, and there are limited single points of failure. An example of a failure that will result in an aborted mission and an emergency ascent is if a leak detection sensor is triggered.

A final active emergency response if the robot is not rising to the surface due to loss of power to achieve this or for any other reason, it will release the load in the hopper; actively attempted if control electronics and power are available or triggered passively using a time delay corrosion fuse that will release the load after approximately 1.5 weeks have passed.

In all cases, a locating beacon on a separate power supply will ping to inform the surface ship of its position if the beacon loses communication with the rest of the vehicle

While no substantially specialized tools are required for maintaining the underwater robots, while onboard the vessel, it is impossible to travel to the hardware store to pick up what is needed, so a complete set of required tools and redundancies will be available on the vessel.

No, the limitation of our operations is the rate at which the underwater robots are lifted onto the vessels. By increasing the number of vessels concurrently operating, the number of underwater robots doesn’t have limitations.

Many terabytes of storage will be available onboard the underwater robots, providing more data storage capacity than is required. Additionally, redundancy will be in place to protect against corrupted drive failures.

An acoustic link (USBL) transmits basic telemetry and status information between the underwater robots and the surface ship. Due to the low bandwidth of the acoustic link, we are limited in terms of what data is transmitted while the robots are submerged. Once a robot has surfaced as part of its on-deck servicing, the onboard data is transferred off the robot and into the on-ship data center. Data is exchanged between the ship and the cloud-based servers via TCP/IP connection over satellite.

The mission is downloaded to the underwater robot at the start of the mission. We don’t need real-time communication. If an emergency arrives, the robot will go to the surface and use WiFi or satellite communication to communicate with the operating center.

Temperature and salinity at the depths at which we operate are very consistent. This impacts the index of refraction for our optical and acoustic systems, and variation in these values at the depth and through the water column is compensated for during operation.

A common practice for docking autonomous underwater vehicles is securing them below the water’s surface so surface waves do not impact them and changing weather conditions. We are using this approach for the underwater robot’s recovery operations.

Our techno-economic models assume 40 days a year when we are in a ‘weather hold’ condition and cannot operate.

We design for this using pressure vessels to protect the components and, in other cases, oil-compensate when the internal electronics are pressure tolerant but need to be isolated from the seawater. These are both common techniques for design in underwater robots.

From the extensive exploration data, we know that there are very few currents in the deep ocean. We have thrusters and control surfaces for navigating the underwater robot in the water column. Also, in the mission control software, we receive the sensor data in real-time via USBL and update the mission parameters. 

At the altitude of 1 meter where we are currently operating, the resolution ranges from 1.25 mm to 0.38 mm on the seafloor, depending on which camera and where in the camera field of view is being considered.

The nodule location is a consistent environment with a flat seabed and clear water. The underwater robot does not need to adapt to different environments autonomously. The size and shape of the nodules can vary from one region with nodules to another, and the algorithms for these new environments will need to be adapted. Still, this adaptation takes place at the development level, not in the autonomous behavior of the underwater robot.

The design of the robotic arms, the claw, the arm’s movement, and the claw’s position on the nodule are optimized to minimize local sediment disturbance. Even so, some local sediment disturbance can occur when the nodule is picked. The cameras identify the location of the nodule in front of the vehicle. The nodule’s location is tracked relative to the robot through precise tracking of the vehicle position. Even with the nodule out of sight from the camera, because the robot’s position is precisely tracked, the nodule’s location is understood, enabling the arm to pick it. With the nodule under the vehicle, the arm picks it, and any disturbed sediment is well behind the camera. Additionally, the vehicle will travel primarily into any current that exists. Between the vehicle motion and the surrounding currents, any sediment distributed under the vehicle will remain behind the vehicle.

Each Eureka IV robot has approximately 60 arms with a 25-metric-ton payload. The Eureka IV can be reused every 4 hours. So, in 24 hours, 6 missions can be completed, delivering 6 * 25 = 150 metric tons per 24 hours. A fleet of 128  Eureka IV robots can deliver 19,200 metric tons per 24 hours using 4 vessels, which translates to 6 million metric tons per year, assuming ~312 production days a year. (Assuming around 53 days a year when a weather hold is in operation.)

The U.S. Army Corps of Engineers defines dredging as “A dredge is a machine that scoops or suctions sediment from the bottom of waterways…”

Although companies are using the ‘Coanda effect,’ it is still dredging. The leading dredging companies have built polymetallic nodule collection systems, e.g., All Seas, DEME/GSR, Royal IHC, etc. Dredging tractors for deep sea mining were first described in a 1965 patent, 3,456,371.